International Research Journal of Finance and Economics

ISSN 1450-2887 Issue 14 (2008)

© EuroJournals Publishing, Inc. 2008
http://www.eurojournals.com/finance.htm

Aggregate Economic Variables and

Stock Markets in India

Shahid Ahmed
Economist, India Programme, UNCTAD, New Delhi (India)
Department of Economics, Jamia Millia Islamia (Central University), New Delhi
E-mail: shah_ec_jmi@yahoo.co.in

Abstract

This study investigates the nature of the causal relationships between stock prices
and the key macro economic variables representing real and financial sector of the Indian
economy for the period March, 1995 to March, 2007 using quarterly data. These variables
are the index of industrial production, exports, foreign direct investment, money supply,
exchange rate, interest rate, NSE Nifty and BSE Sensex in India. Johansen`s approach of
cointegration and Toda and Yamamoto Granger causality test have been applied to explore
the long-run relationships while BVAR modeling for variance decomposition and impulse
response functions has been applied to examine short run relationships. The results of the
study reveal differential causal links between aggregate macro economic variables and
stock indices in the long run. However, the revealed causal pattern is similar in both
markets in the short run. The study indicates that stock prices in India lead economic
activity except movement in interest rate. Interest rate seems to lead the stock prices. The
study indicates that Indian stock market seems to be driven not only by actual performance
but also by expected potential performances. The study reveals that the movement of stock
prices is not only the outcome of behaviour of key macro economic variables but it is also
one of the causes of movement in other macro dimension in the economy.

Keywords: Cointegration Test, T-Y Granger Causality Test, Variance Decomposition,

Impulse Response Function
JEL Classification Codes: G1, F4, C22.

Introduction
The movement of stock indices is highly sensitive to the changes in fundamentals of the economy and
to the changes in expectations about future prospects. Expectations are influenced by the micro and
macro fundamentals which may be formed either rationally or adaptively on economic fundamentals,
as well as by many subjective factors which are unpredictable and also non quantifiable. It is assumed
that domestic economic fundamentals play determining role in the performance of stock market.
However, in the globally integrated economy, domestic economic variables are also subject to change
due to the policies adopted and expected to be adopted by other countries or some global events. The
common external factors influencing the stock return would be stock prices in global economy, the
interest rate and the exchange rate. For instance, capital inflows and outflows are not determined by
domestic interest rate only but also by changes in the interest rate by major economies in the world.
Recently, it is observed that contagion from the US sub prime crisis has played significant movement

Electronic
Electroniccopy
copyavailable
availableat:
at:https://ssrn.com/abstract=1693544
http://ssrn.com/abstract=1693544
142 International Research Journal of Finance and Economics - Issue 14 (2008)

in the capital markets across the world as foreign hedge funds unwind their positions in various
markets. Other burning example in India is the appreciation of currency due to higher inflow of foreign
exchange. Rupee appreciation has declined stock prices of major export oriented companies.
Information technology and textile sector are the example of falling stock prices due to rupee
appreciation.
The modern financial theory focuses upon systematic factors as sources of risk and
contemplates that the long run return on an individual asset must reflect the changes in such systematic
factors. This implies that securities market must have a significant relationship with real and financial
sectors of the economy. This relationship generally viewed in two ways. The first relationship views
the stock market as the leading indicator of the economic activity in the country, whereas the second
focuses on the possible impact the stock market may have on aggregate demand, particularly through
aggregate consumption and investment. The former case implies that stock market leads economic
activity, whereas the latter suggests that it lags economic activity. Knowledge of the sensitivity of
stock market to macro economic behaviour of key variables and vice-versa is important in many areas
of investments and finance. This research may be useful to understand this relationship.
From the beginning of the 1990s in India, a number of measures have been taken for economic
liberalization. At the same time, large number of steps has been taken to strengthen the stock market
such as opening of the stock markets to international investors, regulatory power of SEBI, trading in
derivatives, etc. These measures have resulted in significant improvements in the size and depth of
stock markets in India and they are beginning to play their due role. Presently, the movement in stock
market in India is viewed and analysed carefully by large number of global players. Understanding
macro dynamics of Indian stock market may be useful for policy makers, traders and investors. Results
may reveal whether the movement of stock prices is the outcome of something else or it is one of the
causes of movement in other macro dimension in the economy. The study also expects to explore
whether the movement of stock market are associated with real sector of the economy or financial
sector or both. In this context, the objective of this paper is to explore such causal relations for India as
there is a little work for this period. To the best of our knowledge no in-depth such analysis of macro
dynamics is available in the literature for the period of 1995 to 2007.
The paper is organised in the following sections. Section II provides review of selected
literature on the causal relationship between stock prices and macro variables. Section III discusses the
data and explains the methodology for testing the stationarity, the existence of cointegration, and the
direction of causality. Section IV reports the results and their interpretation. Finally, Section V
provides concluding remarks.

II. Review of Selected Literature

Varying evidences of causal links of stock returns and macro variables have been found in the
literature using various asset pricing specifications. In the literature, widely popular CAPM has been
severely challenged since returns can be predicted from other financial factors. This has led to the
development and testing of various alternative asset pricing specifications, such as the Arbitrage
Pricing Theory (APT) and Present Value Model (PVM). In the context of macro dynamics of stock
returns, APT assumes that returns are generated by a number of macroeconomic factors. It allows
multiple risk factors to explain asset returns. Chen, Roll and Ross (1986) have argued that stock returns
should be affected by any factor that influences future cash flows or the discount rate of those cash
flows. In an empirical investigation they found that the yield spread between long and short term
government bonds, expected inflation, unexpected inflation, nominal industrial production growth and
the yield spread between corporate high and low grade bonds significantly explain stock market
returns. An alternative way of linking macroeconomic variables and stock prices is the discounted cash
flow or present value model (PVM). This model relates the stock price to future expected cash flows
and the future discount rate of the cash flows. Again, all macroeconomic factors that influence future
expected cash flows or the discount rate by which the cash flows are discounted should have an

Electronic
Electroniccopy
copyavailable
availableat:
at:https://ssrn.com/abstract=1693544
http://ssrn.com/abstract=1693544
International Research Journal of Finance and Economics - Issue 14 (2008) 143

influence on the stock price. The advantage of the PVM model is that it can be used to focus on the
long run relationship between the stock market and macroeconomic variables.
In the literature, various theoretical reasons have been explained linking behaviour of stock
prices and key macro economic variables. For instance, Friedman (1988) suggests ‘wealth effect and
substitution effect’ as the possible channels through which stock prices might directly effect money
demands in the economy. Friedman (1988) expected that the wealth effect will dominate and thus the
demand for money and stock prices to be positively related. The theoretical basis to examine the link
between stock prices and the real variables are well established in economic literature, e.g., in Baumol
(1965) and Bosworth (1975). The relationship between stock prices and real consumption
expenditures, for instance, is based on the life cycle theory, developed by Ando and Modigliani (1963),
which states that individuals base their consumption decision on their expected life time wealth. Part of
their wealth may be held in the form of stocks linking stock price changes to changes in consumption
expenditure. Similarly, the relationship between stock prices and investment spending is based on the
‘q’ theory of James Tobin (1969), where q is the ratio of total market value of firms to the replacement
cost of their existing capital stock at current prices.
In retrospect of the literature, a number of hypotheses also support the existence of a causal
relation between stock prices and exchange rates. For instance, ‘goods market approaches’ (Dornbusch
and Fischer, 1980) suggest that changes in exchange rates affect the competitiveness of a firm as
fluctuations in exchange rate affects the value of the earnings and cost of its funds as many companies
borrow in foreign currencies to fund their operations and hence its stock price. An alternative
explanation for the relation between exchange rates and stock prices can be provided through ‘portfolio
balance approaches’ that stress the role of capital account transaction. Like all commodities, exchange
rates are determined by market mechanism, i.e., the demand and supply condition. A blooming stock
market would attract capital flows from foreign investors, which may cause an increase in the demand
for a country’s currency. The reverse would happen in case of falling stock prices where the investors
would try to sell their stocks to avoid further losses and would convert their money into foreign
currency to move out of the country. There would be demand for foreign currency in exchange of local
currency and it would lead depreciation of local currency. As a result, rising (declining) stock prices
would lead to an appreciation (depreciation) in exchange rates. Moreover, foreign investment in
domestic equities could increase over time due to benefits of international diversification that foreign
investors would gain. Furthermore, movements in stock prices may influence exchange rates and
money demand because investors’ wealth and liquidity demand could depend on the performance of
the stock market.
Economic theories suggest causal relations between stock prices and exchange rates, existing
evidence also provides relatively stronger relationship between stock price and exchange rate. Ma and
Kao (1990) find that a currency appreciation negatively affects the domestic stock market for an
export-dominant country and positively affects the domestic stock market for an import-dominant
country, which seems to be consistent with the goods market theory. Bahmani and Sohrabian (1992)
found a bi-directional causality between stock prices measured by the Standard & Poor's 500 index and
the effective exchange rate of the dollar, at least in the short run. The co-integration analysis revealed
no long run relationship between the two variables. Similarly, Abdalla and Murinde (1996) investigate
interactions between exchange rates and stock prices in the emerging financial markets of India, Korea,
Pakistan and the Philippines. The results of the granger causality tests results show uni-directional
causality from exchange rates to stock prices in all the sample countries, except the Philippines. Ajayi
and Mougoue (1996), using daily data for eight countries, show significant interactions between
foreign exchange and stock markets, while Abdalla and Murinde (1997) document that a country’s
monthly exchange rates tends to lead its stock prices but not the other way around. Pan, Fok & Lui
(1999) used daily market data to study the causal relationship between stock prices and exchange rates
and found that the exchange rates Granger-cause stock prices with less significant causal relations from
stock prices to exchange rate. They also find that the causal relationship have been stronger after the
Asian crisis.

Electronic copy available at: https://ssrn.com/abstract=1693544

144 International Research Journal of Finance and Economics - Issue 14 (2008)

Chen, Roll, and Ross [1986], Bodie [1976], Fama [1981], Geske and Roll [1983], Pearce and
Roley [1983], Pearce [1985] and many papers have tried to show empirical associations between
macroeconomic variables and security returns. Bodie [1976], Fama [1981], Geske and Roll [1983], and
Pearce and Roley [1983], Pearce [1985] document a negative impact of inflation and money growth on
equity values. Many experts however believe that positive effects will outweigh the negative effects
and stock prices will eventually rise due to growth of money supply (e.g., Mukherjee and Naka, 1995).
They argue that a change in the money supply provides information on money demand, which is
caused by future output expectations. If the money supply increases, it means that money demand is
increasing, which, in effect, signals an increase in economic activity. Higher economic activity implies
higher cash flows, which causes stock prices to rise. Bernanke and Kuttner (2005) argue that the price
of a stock is a function of its monetary value and the perceived risk in holding the stock. A stock is
attractive if the monetary value it bears is high. On the other hand, a stock is unattractive if the
perceived risk is high. The authors argue that the money supply affects the stock market through its
effect on both the monetary value and the perceived risk. Money supply affects the monetary value of a
stock through its effect on the interest rate. The authors believe that tightening the money supply raises
the real interest rate. An increase in the interest rate would in turn raise the discount rate, which would
decrease the value of the stock as argued by the real activity theorists.
The impact of real sector macro variables on equity returns has been much more difficult to
establish. Mukherjee and Naka (1995) reveal that cointegration relation existed and positive
relationship was found between the Japanese industrial production and stock return. However, Cutler,
Poterba, and Summers [1989] (CPS) find that Industrial Production growth is significantly positively
correlated with real stock returns over the period 1926-1986, but not in the 1946- 85 sub-period. In
Indian context, Bhattacharya and Mukherjee (2002) studied the nature of the causal relationship
between stock prices and macro aggregates for the period of 1992-93 to 2000- 2001.Their results show
that there is no causal relationship between stock price and macro economic variables like money
supply, national income and interest rate but there exists a two way causation between stock price and
rate of inflation. Their results also indicate index of industrial production lead the stock price.
As discussed above, literature reveals differential causal pattern between key macro economic
variables and stock prices. This relationship varies in a number of different stock markets and time
horizons in the literature. This paper will add to the existing literature by providing robust result, which
is based on more than one technique, about causal links for the longer period, i.e., 12 years quarterly
data.

III. Empirical Methodology and Data

Time series analysis must be based on stationary data series for drawing useful inferences. Broadly
speaking a data series is said to be stationary if its mean and variance are constant (non-changing) over
time and the value of covariance between two time periods depends only on the distance or lag
between the two time periods and not on the actual time at which the covariance is computed. The
correlation between a series and its lagged values are assumed to depend only on the length of the lag
and not when the series started. This property is known as stationarity and any series obeying this is
called a stationary time series.
Three unit root tests have been applied to test whether a series is stationary or not. Stationarity
condition has been tested using Augmented Dickey Fuller (ADF) and Phillips – Perron (PP) tests.
[Dickey and Fuller (1979, 1981), Gujarati (2003), Phillips and Perron (1988), Enders (1995)]. For
testing null hypothesis of stationarity, KPSS test has also been applied for robustness [Kwiatkowski,
Phillips, Schmidt. And Shin(1992)].

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 145

Augmented Dickey Fuller Test

Augmented Dickey-Fuller (ADF) test has been carried out which is the modified version of Dickey –
Fuller (DF) test. ADF makes a parametric correction in the original DF test for higher-order correlation
by assuming that the series follows an AR (p) process. The ADF approach controls for higher-order
correlation by adding lagged difference terms of the dependent variable to the right-hand side of the
regression. The Augmented Dickey-Fuller test specification used here is as follows:
p
yt 0 yt 1 i yt i ut
i 1 (1)

Phillips-Perron (PP) Test

Phillips and Perron (1988) adopts a nonparametric method for controlling higher-order serial
correlation in a series. The test regression for the Phillips-Perron (PP) test is the AR (1) process. While
the ADF test corrects for higher order serial correlation by adding lagged differenced terms on the
right-hand side, the PP test makes a correction to the t-statistic of the coefficient from the AR(1)
regression to account for the serial correlation in ut . The correction is nonparametric. The advantage of
Phillips-Perron test is that it is free from parametric errors. Phillips-Perron (PP) test allows the
disturbances to be weakly dependent and heterogeneously distributed. In view of this, PP values have
also been checked for stationarity.

KPSS Test
A major criticism of the ADF unit root testing procedure is that it cannot distinguish between unit root
and near unit root processes especially when using short samples of data. This prompted the use of the
KPSS test, where the null is of stationarity against the alternative of a unit root. This ensures that the
alternative will be accepted (null rejected) only when there is strong evidence for (against) it
[Kwiatkowski, Phillips, Schmidt. and Shin (1992)].

Co-integration Test
Using non-stationary series, cointegration analysis has been used to examine whether there is any long-
run equilibrium relationship. For instance, when non-stationary series are used in regression analysis,
one as a dependent variable and the other as an independent variable, statistical inference becomes
problematic [Granger and Newbold, 1974]. Cointegration analysis becomes important for the
estimation of error correction models (ECM). The concept of error correction refers to the adjustment
process between short-run disequilibrium and a desired long run position. As Engle and Granger
(1987) have shown, if two variables are cointegrated, then there exists an error correction data
generating mechanism, and vice versa. Since, two variables that are cointegrated, would on average,
not drift apart over time, this concept provides insight into the long-run relationship between the two
variables and testing for the cointegration between two variables. In the present case, Johansen’s
Maximum Likelihood procedure for Cointegration has been applied.
Johansen (1991) method can be illustrated by considering the following general autoregressive
representation for the vector Y.
p
Yt A0 AjYt j
j 1
(2)
where Yt is an n 1 vector of non stationary I(1) variables, A0 is an n 1 vector of constants, p is the
number of lags, Aj is a (n x n) matrix of coefficients and t is assumed to be a ( n 1 ) vector of
Gaussian error terms.
In order to use Johansen’s test, the above vector autoregressive process can be reparametarized
and turned into a vector error correction model of the form:

Electronic copy available at: https://ssrn.com/abstract=1693544

146 International Research Journal of Finance and Economics - Issue 14 (2008)
p 1
Yt A0 j Yt j Yt p t
j 1
(3)
Where,
p

j Aj
i j 1

and
p
I Aj
i j 1

is the difference operator, and I is an (n x n) identity matrix.

The issue of potential co-integration is investigated by comparing both sides of equation (4). As
Yt ~ I(1) , Yt ~ I(0) , so are Yt-j . This implies that the left-hand side of equation (4) is stationary.
Since Yt-j is stationary, the right-hand side of equation (4) will also be stationary if Yt-p is
stationary. The Johansen test centers on an examination of the matrix. The can be interpreted as
a long run coefficient matrix, since in equilibrium, all the Yt-j will be zero, and setting the error terms,
t, to their expected value of zero will leave Yt-p = 0. The test for co-integration between the Y’s is
calculated by looking at the rank of the matrix via eigenvalues. The rank of a matrix is equal to the
number of its characteristic roots (eigenvalues) that are different from zero. There are three possible
cases to be considered: Rank ( ) = p and therefore vector Xt is stationary; Rank ( ) = 0 implying
absence of any stationary long run relationship among the variables of Xt or Rank ( ) < p and
therefore r determines the number of cointegrating relationships. Thus, if the rank of equals to 0, the
matrix is null and equation (4) becomes the usual VAR model in first difference. If the rank of is r
where r < n, then there exist r co-integrating relationships in the above model. In this case, the matrix
'
can be rewritten as = where and are n r matrices of rank r. Here, is the matrix
of co-integrating parameters and is the matrix of weights with which each co-integrating vector
enters the above BVAR model.
The test for the number of characteristic roots that are insignificantly different from unity can
be conducted using the following two statistics, namely, the trace and the maximum eigenvalue test.
p
ˆ)
trace ( r ) T ln(1 j
j r 1
(4)
and
(r , r 1) T ln(1 ˆ )
max r 1 (5)
ˆ
Where j is the estimated values of the characteristic roots (also called eigenvalue) obtained
from the estimated matrix, T is the number of usable observations. r is the number of co-integrating
vectors.
The trace test statistics test the null hypothesis that the number of distinct co-integrating vectors
is less than or equal to r against the alternative hypothesis of more than r co-integrating relationships.
From the above it is clear that trace equals zero when all ˆ j =0. The farther the estimated
characteristic roots are from zero, the more negative is ln(1- ˆ ) and larger the j statistics. The
trace

maximum eigenvalue statistics test the null hypothesis that the number of co-integrating vectors is less
than or equal to r against the alternative of r +1 co-integrating vectors. Again, if the estimated value of
the characteristic root is close to zero, max will be small (Enders, 1995; Madala and Kim, 1998).

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 147

Toda and Yamamoto (T-Y) Granger Causality Test

The dynamic linkage is examined using the concept of Granger’s causality test (1969, 1988). A time
series xt Granger-causes another time series yt if series yt can be predicted with better accuracy by
using past values of xt rather than by not doing so, other information is being identical. In other words,
variable xt fails to Granger-cause yt if
Pr( y t+m ) =Pr( y t+m )
t t
(6)
Where Pr( y t+m t ) denotes conditional probability of yt , where t is the set of all
information available at time t, and Pr( y t+m t ) denotes conditional probability of yt obtained by
excluding all information on xt from yt . This set of information is depicted as t . The causal linkage
between stock prices and macro economic aggregates in India are investigated by applying the
technique of long run Granger non causality test developed by T-Y(1995).
The selection of the VAR system requires an analysis of unit roots and cointegration which
may cause inadequate results 1 (Blough, 1992). This can lead us to select an incorrect model for
verifying the relations of causality, possibly causing a problem of over-rejection of non causal null
hypothesis (Giles and Mirza, 1999). In this way, Toda and Yamamoto (1995), Dolado and Luketerpohl
(1996) propose an applicable methodology independent of the integration or co-integration properties
of the model. In this method a modified Wald Test is used to contrast the parameters of the VAR. An
extended VAR model is used, whose order is determined by the number of optimal lag lengths in the
system (k) and the maximum number of times one must differentiate the variables (d max ) . When a
VAR (k d max ) is predicted (where d max is the maximum order of integration to occur in the system),
this test displays asymptotic chi-square distribution, it is also shown that if variables are integrated of
order d, the usual selection procedure is valid whenever k d. Toda and Yamamoto test has been used
to capture long-run causality pattern of stock indices and the following specification has been used to
estimate:
k d max k d max
Xt 1 11 i Xt ( k d max ) 12 j Yt ( k dmax ) Xt (7)
i 1 j 1
k d max k d max
Yt 2 21 i Xt ( k dmax ) 22 j Yt ( k dmax ) Yt (8)
i 1 j 1

where Yt and xt are stationary random processes intended to capture other pertinent information not
contained in lagged values of X t and Yt . The lag length k is decided by AIC in our study. The series
Yt fails to Granger cause X t if 12(j)= 0 (j=1,2,3, k d max ); and the series X t fails to cause Yt . If
21(i)= 0 (i=1,2,3,.. k d max ).

Variance decomposition
The vector autoregression (VAR) by Sims (1980) has been estimated to capture short run causality
between stock prices and key economic macro economic variables. Variance decomposition and
impulse response function has been utilized for drawing inferences. Equation system is similar to the
equations for granger causality assuming simultaneity among a set of variables and treating all the
variables endogenous to system. VAR is commonly used for forecasting systems of interrelated time
series and for analyzing the dynamic impact of random disturbances on the system of variables. The
VAR approach sidesteps the need for structural modeling by treating every endogenous variable in the
system as a function of the lagged values of all of the endogenous variables in the system. In the
present study, BVAR model has been specified in first differences as given in equation 10 and 11.

1
The works of Bewley and Yang (1996) and Ho and Sorensen (1996) illustrate the problems associated with the utilization of the Johansen and
Juselius test on the analysis of cointegration

Electronic copy available at: https://ssrn.com/abstract=1693544

148 International Research Journal of Finance and Economics - Issue 14 (2008)
k k
Xt 1 11 i Xt i 12 j Yt j Xt (9)
i 1 j 1
k k2
Yt 2 21 i Xt i 22 j Yt j Yt (10)
i 1 j 1

Where ’s are the stochastic error terms, called impulse response or innovations or shock in
the language of VAR. Impulse response functions have been estimated to trace the effects of a shock to
one endogenous variable on to the other variables in the BVAR. The impulse response functions can be
used to produce the time path of the dependent variables in the BVAR, to shocks from all the
explanatory variables. If the system of equations is stable any shock should decline to zero, an unstable
system would produce an explosive time path.
Variance decomposition (Choleski Decomposition) is the alternative way which separates the
variation in an endogenous variable into the component shocks to the BVAR. Thus, the variance
decomposition which provides information about the relative importance of each random innovation in
affecting the variables in the BVAR has also been presented. In econometric literature, both impulse
response functions and variance decomposition together are known as innovation accounting (Enders,
1995).
For empirical investigation, the data set for the India consists of the quarterly NSE Nifty, BSE
Sensex, Index of Industrial Production, Money Supply, Interest Rate, Foreign Direct Investment
Inflows and Export Earnings for the period March, 1995 to March 2007. The Indian capital market is
represented by the National Stock Exchange (NSE) and the Bombay Stock Exchange (BSE). The data
for the NSE Nifty and BSE Sensex has been taken from their respective websites. The data for other
macro key variables have taken from Ministry of Statistics and Programme Implementation, Govt. of
India and Economic Intelligence Service, Centre for Monitoring Indian Economy Pvt. Ltd.

IV. Empirical Analysis

The descriptive statistics for all eight variables are calculated and presented in table 1. These variables
are Export earnings, Money Supply, Exchange Rate, Interest Rate, Index of Industrial Production,
Foreign Direct Investment, NSE Nifty and BSE Sensex. The skewness coefficient, in excess of unity is
taken to be fairly extreme [Chou 1969]. High or low kurtosis value indicates extreme leptokurtic or
extreme platykurtic [Parkinson 1987]. Generally values for zero skewness and kurtosis at 3 represents
that the observed distribution is normally distributed. It is seen that the frequency distribution of the
above mentioned variables are not normal. Jarque-Bera statistics also indicates that the frequency
distribution of the underlying series does not fit normal distribution. Further, the standard deviation
indicates that the Export earnings, Money Supply, FDI, NSE Nifty and BSE Sensex are relatively more
volatile compared to Exchange Rate, Interest Rate and Index of Industrial Production.

Table 1: Descriptive Statistics

Variables EXPORT EXRATE IIP INT MS FDI NIFTY SENSEX

Mean 8.33 3.75 5.12 2.40 14.06 6.76 7.22 8.41
Median 8.20 3.78 5.09 2.35 14.09 6.72 7.04 8.24
Maximum 9.44 3.89 5.65 2.72 15.01 10.03 8.31 9.55
Minimum 7.71 3.45 4.73 2.20 13.18 3.89 6.73 7.94
Std. Dev. 0.47 0.12 0.21 0.17 0.52 0.80 0.41 0.42
Skewness 0.73 -1.02 0.28 0.46 -0.06 0.72 1.15 1.24
Kurtosis 2.28 2.85 2.30 2.05 1.83 5.31 3.31 3.55
Jarque-Bera 16.15 25.17 4.90 10.68 8.32 44.92 32.81 39.31
Probability 0.00 0.00 0.09 0.00 0.02 0.00 0.00 0.00

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 149

The first and simplest type of test one can apply to check for stationarity is to actually plot the
time series and look for evidence of trend in mean, variance, autocorrelation and seasonality. If any
such patterns are present then these are signs of non-stationarity. The eight time series displayed in
figure-1 exhibit different such patterns. Export, IIP and Money Supply seems to exhibit a trend in the
mean since it has a clear upward slope. In fact, sustained upward or downward sloping patterns (linear
or non-linear) are signs of a non-constant mean. The time series on Exchange Rate, FDI, NSE Nifty
and BSE sensex in the figure contains an obvious trend in both mean and variance. Along with a trend
in the mean, the vertical fluctuation of the series also appears to differ greatly from one portion of the
series to the other, indicating that the variance is not constant. Once again this is a sign of non-
stationarity.

Electronic copy available at: https://ssrn.com/abstract=1693544

150 International Research Journal of Finance and Economics - Issue 14 (2008)
Figure 1: Dataset Graph

9 .6
8 .4

9 .2
8 .0

8 .8
7 .6

8 .4
7 .2

8 .0
6 .8

7 .6
6 .4
95 96 97 98 99 00 01 02 03 04 05 06
95 96 97 98 99 00 01 02 03 04 05 06
SENSEX
N IF T Y

9 .6 3 .9

3 .8
9 .2

3 .7
8 .8

3 .6
8 .4

3 .5
8 .0
3 .4
95 96 97 98 99 00 01 02 03 04 05 06
7 .6
95 96 97 98 99 00 01 02 03 04 05 06 E X R A TE

E X P ORT
.
15 .2
2.8
14 .8
2.7
14 .4
2.6
14 .0
2.5
13 .6
2.4

2.3 13 .2

2.2 12 .8
95 96 97 98 99 00 01 02 03 04 05 06
2.1
95 96 97 98 99 00 01 02 03 04 05 06 MS

INT

11
5 .8
10
5 .6 9

5 .4 8

5 .2 7

6
5 .0
5
4 .8
4
4 .6
95 96 97 98 99 00 01 02 03 04 05 06 3
95 96 97 98 99 00 01 02 03 04 05 06
IIP
FDI

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 151
Figure 2: Impulse Response Functions-Nifty and Other Macro Variables

R e s p o n s e o f D (N IF T Y ) to C h o le s k y
O n e S .D . D ( E X R A T E ) In n o va ti o n
R e s p o n s e o f D ( E X R A T E ) t o C h o le s k y
O n e S . D . D ( N IF T Y ) In n o v a t i o n .0 0 5
.0 0 3
.0 0 0
.0 0 2

.0 0 1 -.0 0 5
.0 0 0
-.0 1 0
- .0 0 1

- .0 0 2
-.0 1 5
- .0 0 3

- .0 0 4
-.0 2 0

- .0 0 5
-.0 2 5
1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10

R e s p o n s e o f D ( E X P O R T ) to C h o le s k y
O n e S . D . D ( N IF T Y ) In n o v a t i o n
.0 6 R e s p o n s e o f D ( N IF T Y ) to C h o le s k y
O n e S . D . D ( E X P O R T ) In n o v a ti o n

.0 4 .0 2 0
.0 1 6
.0 2 .0 1 2
.0 0 8
.0 0
.0 0 4
- .0 2 .0 0 0

- .0 0 4
- .0 4
- .0 0 8

- .0 6 - .0 1 2
1 2 3 4 5 6 7 8 9 10 - .0 1 6
1 2 3 4 5 6 7 8 9 10

R e s p o n s e o f D ( IIP ) t o C h o le s k y
O n e S . D . D ( N IF T Y ) In n o v a t i o n
R e s p o n s e o f D ( N IF T Y ) to C h o le s k y
.0 2
O n e S .D . D ( IIP ) In n o v a t i o n
.0 2 4
.0 2 0 .0 1
.0 1 6
.0 1 2
.0 0
.0 0 8
.0 0 4
-.0 1
.0 0 0
- .0 0 4
- .0 0 8 -.0 2
1 2 3 4 5 6 7 8 9 10
- .0 1 2
1 2 3 4 5 6 7 8 9 10

R e s p o n s e o f D (M S ) to C h o le s k y R e s p o ns e o f D (N IF TY) to C ho le s k y
O n e S . D . D ( N I F T Y ) I n n o v a ti o n O ne S .D . D (M S ) Inno va ti o n
.0 0 15 .0 2 5

.0 0 10 .0 2 0

.0 0 05 .0 1 5

.0 0 00 .0 1 0

-.0005 .0 0 5

-.0010 .0 0 0

-.0015 -.0 0 5

-.0020 -.0 1 0
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Electronic copy available at: https://ssrn.com/abstract=1693544

152 International Research Journal of Finance and Economics - Issue 14 (2008)

Response of D(NIFTY) to Cholesky Response of D(INT) to Cholesky

One S.D. D(INT) Innovation One S.D. D(NIFTY) Innovation
.004
.03
.003
.02 .002

.001
.01
.000

.00 -.001

-.002
-.01 -.003
-.004
-.02
-.005
1 2 3 4 5 6 7 8 9 10
-.03
1 2 3 4 5 6 7 8 9 10

R e s p o n s e o f D (N IF T Y) to C h o le s k y R e s p o ns e o f D (F D I) to C ho le s k y
O n e S .D . D ( F D I) In n o v a tio n O ne S .D . D (N IF TY) Inno va tio n
.0 3 .2

.0 2
.1

.0 1

.0
.0 0

-.1
-.0 1

-.0 2 -.2
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

R e s p o n s e o f D ( S E N S E X ) t o C h o le s k y
R e s p o n s e o f D ( N IF T Y ) to C h o le s k y O n e S .D . D ( N IF T Y ) In n o v a ti o n
O n e S .D . D ( S E N S E X ) In n o v a ti o n
.0 3
.1 0

.0 8 .0 2

.0 6 .0 1

.0 4 .0 0

.0 2
- .0 1
.0 0
- .0 2
- .0 2
- .0 3
- .0 4 1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 153
Figure 3: Impulse Response Functions-Sensex and Other Macro Variables

Re sp o nse o f D (S E N S E X ) to C ho le s ky
Re sp o nse o f D (E X RA TE ) to C ho le sky O ne S .D . D (E X R A TE ) Inno va tio n
O ne S .D . D (S E NS E X ) Inno va tio n .0 1 0
.0 01
.0 0 5
.0 00
.0 0 0
-.00 1
-.0 0 5
-.00 2
-.0 1 0
-.00 3
-.0 1 5
-.00 4
-.0 2 0
-.00 5
-.0 2 5
1 2 3 4 5 6 7 8 9 10
-.00 6
1 2 3 4 5 6 7 8 9 10

R e sp o nse o f D (E X P O R T) to C hole s ky
O ne S .D . D (S E N S E X ) Inno va tio n
.04
R e s p o ns e o f D (S E N S E X ) to C ho le s k y
O ne S .D . D (E X P O R T) Inno va ti o n
.03
.0 2 0
.02
.0 1 5

.0 1 0 .01

.0 0 5 .00
.0 0 0
-.01
-.0 0 5
-.02
-.0 1 0

-.0 1 5 -.03
1 2 3 4 5 6 7 8 9 10
-.0 2 0
1 2 3 4 5 6 7 8 9 10

Response of D (IIP ) to C holesky

One S .D . D (S E NS E X ) Innovation R e s p o n s e o f D ( S E N S E X ) to C h o le s k y
.015 O n e S .D . D ( IIP ) In n o v a ti o n
.0 2 0
.010
.0 1 5

.005 .0 1 0

.0 0 5
.000
.0 0 0
-.005
- .0 0 5
-.010
- .0 1 0

-.015 - .0 1 5
1 2 3 4 5 6 7 8 9 10
-.020
1 2 3 4 5 6 7 8 9 10

R e s p o ns e o f D ( IN T ) to C ho le s k y R e s p o n s e o f D ( S E N S E X ) to C h o le s k y
O ne S .D . D (S E N S E X ) Inn o v a ti o n O n e S .D . D ( IN T ) In n o v a ti o n

.0 0 4 .0 3

.0 2
.0 0 2
.0 1

.0 0 0
.0 0

-.0 0 2 - .0 1

- .0 2
-.0 0 4
- .0 3
1 2 3 4 5 6 7 8 9 10
-.0 0 6
1 2 3 4 5 6 7 8 9 10

Electronic copy available at: https://ssrn.com/abstract=1693544

154 International Research Journal of Finance and Economics - Issue 14 (2008)
Response of D (MS ) to C holesky Response of D (SE NS E X ) to C holesky
One S .D . D (S E NS E X ) Innovation One S .D . D (MS ) Inno vation
.003 .010

.002 .005

.001 .000

.000 -.005

-.001 -.010

-.002 -.015

-.003 -.020
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

Response of D(SENSEX) to Cholesky Response of D(FDI) to Cholesky

One S.D. D(FDI) Innovation One S.D. D (SENSEX) Innovation
.03 .2

.02 .1

.01
.0

.00
-.1
-.01

-.2
-.02 1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10

As indicated by visual inspection, formal test for stationarity is essential to opt for appropriate
methodological framework. As a first step, we tested the eight variables (Exports, Exchange Rate,
Index of Industrial Production, Interest Rate, Money Supply, FDI, Nifty and Sensex) for stationarity by
applying ADF, PP unit root test and KPSS stationarity test. ADF, PP and KPSS statistics are given in
Table-2. On the basis of ADF statistics and PP test, all the series are found to be non-stationary at
levels except exchange rate. In the series of exchange rate, ADF and PP test statistics rejects null
hypotheses of unit root in case of levels at 5 percent and 6 percent level of significance. Further, ADF
statistics rejects null hypotheses of unit root in case of first differences of money supply only at 10
percent. Finally, KPSS test is applied which has null of stationarity. In this case, all variables including
exchange rate are non stationary in levels and stationary in first differences. Assuming all the variables
are non-stationary at levels and stationary at first differences on the basis of KPSS test and visual
inspections, Johansen’s approach of cointegration, Toda and Yamamoto Granger causality test and
VAR modeling for variance decomposition/impulse response functions have been applied.

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 155
Table 2: Unit Root Test

Null Hypothesis: Variable Null Hypothesis: Variable Null Hypothesis: Variable

is Non-stationary is Non-stationary is stationary
Variables Augmented Dickey-Fuller Phillips-Perron Test Kwiatkowski-Phillips-
test statistic Statistics Schmidt-Shin test statistic
Level First Difference Level First Difference Level First Difference
Exports 2.43 -2.93* -0.07 -27.40* 1.31* 0.39
Exchange Rate -2.88** -9.04** -2.86*** -8.98* 1.02* 0.65
Index of Industrial Production 3.13 -2.71** 1.01 -23.69* 1.41* 0.37
FDI 0.91 -9.99* 0.09 -55.39* 0.68** 0.14
Interest Rate -1.10 -9.06* -1.18 -8.85* 1.32* 0.15
Money Supply -0.03 -2.68*** 0.93 -10.59* 1.42* 0.15
Nifty 0.14 -13.10* 0.19 -13.07* 1.06* 0.29
Sensex 0.37 -12.56* 0.33 -12.56* 0.95* 0.31
Asymptotic critical values*:
1% level -3.48 -3.47 0.74
5% level -2.88 -2.88 0.46
10% level -2.57 -2.57 0.35
* implies significant at 1% level, ** implies significant at 5% level and *** implies significant at 10% level

To explore whether there is any long-run relationship between Indian stock markets and macro
economic variables such as exports, exchange rate, index of industrial production, foreign direct
investment, interest rate and money supply, Johansen’s cointegration test has been applied. The
number of lags in cointegartion analysis is chosen on the basis of Akaike Information Criteria. Before
discussing the results, it is important to discuss what it implies when two variables are cointegrated and
when they are not. When two variables are cointegrated, it implies that the two time series cannot
wander off in opposite directions for very long without coming back to a mean distance eventually. But
it doesn’t mean that on a daily basis the two series have to move in synchrony at all. When two series
are not cointegrated it implies that the two time series can wander off in opposite directions for very
long without coming back to a mean distance eventually. Results indicate that NSE Nifty-Index of
Industrial production and Money Supply- NSE Nifty may be cointegrated in the long run as the results
vary depending on the varying assumption about trend and intercept. However, FDI- NSE Nifty is
cointegrated in the long run under all assumptions. In case of Exchange Rate-NSE Nifty; Export- NSE
Nifty and Interest Rate- NSE Nifty, there is no evidence of co-integration. Similar test was conducted
between Sensex and above referred macro variables. The results remain the same except that the study
could not reveal any long run relationship between FDI and Sensex as reported in case of NSE and
Nifty case (see Table-3 and 4).

Electronic copy available at: https://ssrn.com/abstract=1693544

156 International Research Journal of Finance and Economics - Issue 14 (2008)
Table 3: Johansen Co-Integration Test: Nifty and Other Macro Variables (Number of Cointegrating Relations
by Model)

Data Trend: None None Linear Linear Quadratic

Test Type No Intercept Intercept Intercept Intercept Intercept
No Trend No Trend No Trend Trend Trend
Nifty- Exchange Rate(3)
Trace 0 0 0 0 0
Max-Eig 0 0 0 0 0
Nifty- Exports(3)
Trace 2 0 0 0 0
Max-Eig 0 0 0 0 0
Nifty- Index of Industrial Production(2)
Trace 0 0 0 1 1
Max-Eig 0 0 0 1 1
Nifty- Interest Rate(4)
Trace 0 0 0 0 0
Max-Eig 0 0 0 0 0
Nifty- Money Supply(1)
Trace 1 1 0 0 0
Max-Eig 1 1 0 0 0
Nifty- Foreign Direct Investment(2)
Trace 1 1 1 1 1
Max-Eig 1 1 1 1 1
Nifty-Sensex(3)
Trace 0 0 0 0 0
Max-Eig 0 0 0 0 0
Selected (0.05 level*) Appropriate lag is given in parantheses on the basis of AIC

To test long run direction of causality, Toda and Yamamoto Granger causality test was
conducted between Nifty and other variables. The results are presented in Table-4. In case of exchange
rate-NSE Nifty and IIP-NSE Nifty, there is no causality in either direction. The study reveals that
export performance does not have any effect on NSE Nifty while NSE Nifty affects export flows from
India. Further, results reveal that Interest Rate causes NSE Nifty while there is no causality from NSE
Nifty to Interest Rate. In case of Money Supply and NSE Nifty, there is no direct causal links. Money
supply might be affecting Nifty through its effect on interest rate. In case of FDI-NSE Nifty, FDI does
cause Nifty movement while movement in NSE Nifty does not have significant effect on IIP. This
links may be the outcome of positive effect of FDI inflows on expectations regarding growth, earning
and profit prospect of the sector. In case of Nifty and Sensex, Sensex causes Nifty while movement in
NSE Nifty does not have significant effect on Sensex.

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 157
Table 4: Johansen Co-Integration Test: Sensex and Other Macro Variables (Number of Cointegrating
Relations by Model)

Data Trend: None None Linear Linear Quadratic

Test Type No Intercept Intercept Intercept Intercept Intercept
No Trend No Trend No Trend Trend Trend
Sensex- Exchange Rate(2)
Trace 0 1 0 0 0
Max-Eig 0 1 0 0 0
Sensex- Exports(3)
Trace 1 0 0 0 0
Max-Eig 0 0 0 0 0
Sensex- Index of Industrial Production(3)
Trace 1 0 0 1 1
Max-Eig 0 0 0 1 1
Sensex- Interest Rate(4)
Trace 0 0 0 0 0
Max-Eig 0 0 0 0 0
Sensex- Money Supply(1)
Trace 1 1 0 0 0
Max-Eig 1 1 0 0 0
Nifty- Foreign Direct Investment(4)
Trace 0 0 0 0 0
Max-Eig 0 0 0 0 0
Selected (0.05 level*) Appropriate lag is given in parantheses on the basis of AIC

Similar Toda and Yamamoto test for causality were conducted between BSE Sensex and other
macro economic variables. The results are presented in Table-5. The results reveal that exchange rate
does not influence the BSE Sensex while movement in BSE Sensex does cause change in exchange
rate. There is no evidence that export performance does have any effect on BSE Sensex while BSE
Sensex causes export flows from India. It is found that Index of Industrial production (IIP) does not
have any influence on Sensex movement while movement in BSE Sensex has significant effect on IIP.
Results reveal that Interest Rate causes BSE Sensex while there is no causality from BSE Sensex to
Interest Rate. In case of BSE Sensex -Money Supply and BSE Sensex -FDI, there is no causal link
observed in ether direction.

Table 5: T&Y(1995) Test of Granger Causality: Nifty and Other Macro Variables.

Wald Test: Chi-

Cause Effect Causality Inference Model
square Values
Exchange Rate Nifty 5.24(3+1) Exchange Rate does not causes Nifty VAR(k +dmax)
Nifty Exchange Rate 0.92(3+1) Nifty does not causes Exchange Rate VAR(k +dmax)
Export Nifty 0.56(3+1) Export does not causes Nifty VAR(k +dmax)
Nifty Export 7.38**(3+1) Nifty causes Export VAR(k +dmax)
Index Of Industrial
Nifty 3.66(2+1) IIP does not cause Nifty VAR(k +dmax)
Production(IIP)
Index of Industrial
Nifty 2.52(2+1) Nifty does not cause IIP VAR(k +dmax)
Production(IIP)
FDI Nifty 5.01***(2+1) FDI cause Nifty VAR(k +dmax)
Nifty FDI 0.92(2+1) Nifty does not cause FDI VAR(k +dmax)
Interest Rate Nifty 9.54*(4+1) Interest Rate causes Nifty VAR(k +dmax)
Nifty Interest Rate 4.88(4+1) Nifty does not cause Interest Rate VAR(k +dmax)
Money Supply Nifty 0.94(1+1) Money Supply does not cause Nifty VAR(k +dmax)
Nifty Money Supply 0.03(1+1) Nifty causes Money Supply VAR(k +dmax)
Nifty Sensex 1.86(3+1) Nifty does not cause Sensex VAR(k +dmax)
Sensex Nifty 1382.92*(3+1) Sensex causes Nifty VAR(k +dmax)
* implies significant at 1% level, ** implies significant at 5% level and *** implies significant at 10% level.

Electronic copy available at: https://ssrn.com/abstract=1693544

158 International Research Journal of Finance and Economics - Issue 14 (2008)

In the context of varying causal links of both stock indices with macro variables, Sims’s VAR
were applied and short run causal links were explored using Variance decomposition and Impulse
response functions. Variance decomposition determines how much of the k-step ahead forecast error
variance of given variable is explained by innovations to each explanatory variable. In practice, it is
usually observed that own series shocks most of the (forecast) error variance of the series in the VAR.
Variance decomposition separates the variation in an endogenous variable into the component shocks
to the VAR and provides information about the relative importance of each random innovation in
affecting the variables in the VAR. The variance decomposition results at the end of 10 periods are
shown in Table 7. The columns provide the percentage of the forecast variance due to each innovation
in bivariate VAR (BVAR) framework, with each row adding up to 100. In case of bivariate modeling
of Nifty and exchange rate, NSE Nifty explains nearly 97% of its own forecast error variance while
Exchange rate explains only 3 percent of NSE Nifty. Exchange Rate explains over 91% of its own
variance and Nifty explains remaining 9 percent of variance of exchange rate. It implies that NSE Nifty
causes Exchange Rate. Variance decomposition in case of Nifty-Export BVAR Model indicates that
NSE Nifty explains almost 100 percent of its variance while NSE nifty explains about 7 percent of
variance of exports. It is an indication that NSE Nifty causes exports. In case of Nifty and IIP BVAR,
Nifty explains 23 percent of variance of IIP while IIP explains about 20 percent of variance in NSE
Nifty. In this case, Nifty leads IIP. In case of Nifty-Interest Rate BVAR, results indicate that interest
rate explains 7.29 percent variance in NSE Nifty while NSE Nifty explains 3.72 percent of Interest
rate, thus Interest Rate affects NSE Nifty. In case of Nifty and FDI BVAR, FDI explains more than 4
percent of variance of Nifty while Nifty explains only 2 percent of FDI variance. FDI seems to have
some lead to NSE Nifty. In case of Nifty and Sensex BVAR, Sensex explains 90 percent of variance of
Nifty while Nifty explains about 3 percent of variance in Sensex. In this case, it is quite clear that
Sensex causes Nifty. Money supply only explains 2 percent of NSE Nifty while 5 percent of variance
of Money supply is explained by NSE Nifty. Direction of effect seems from NSE Nifty to Money
Supply.

Table 6: T&Y(1995) Test of Granger Causality:Sensex and Other Macro Variables

Wald Test: Chi-

Cause Effect Causality Inference Model
square Values
Exchange Rate Sensex 1.92(2+1) Excahnge Rate does not causes VAR(k +dmax)
Sensex
Sensex Exchange Rate 8.35*(2+1) Sensex causes Exchange Rate VAR(k +dmax)
Export Sensex 1.38(3+1) Export does not causes Sensex VAR(k +dmax)
Sensex Export 6.24**(3+1) Sensex causes Export VAR(k +dmax)
Index Of Industrial Sensex 3.06(3+1) IIP does not cause Sensex VAR(k +dmax)
Production(IIP)
Sensex Index of Industrial 8.14*(3+1) Sensex cause IIP VAR(k +dmax)
Production(IIP)
FDI Sensex 4.21(4+1) FDI does not cause Sensex VAR(k +dmax)
Sensex FDI 4.58(4+1) Sensex does not cause FDI VAR(k +dmax)
Interest Rate Sensex 10.76*(4+1) Interest Rate causes Sensex VAR(k +dmax)
Sensex Interest Rate 5.37(4+1) Sensex does not cause Interest Rate VAR(k +dmax)
Money Supply Sensex 0.37(1+1) Money Supply does not cause Sensex VAR(k +dmax)
Sensex Money Supply 1.00(1+1) Sensex causes Money Supply VAR(k +dmax)
* implies significant at 1% level, ** implies significant at 5% level and *** implies significant at 10% level.

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 159
Table 7: VARIANCE DECOMPOSITION OF NIFTY AND OTHER VARIABLES (Bivariate VAR
Framework)

Variance Decomposition of Lags: Nifty Exchange Rate

2 98.58 1.41
5 96.06 3.93
Nifty 10 96.04 3.95
2 4.99 95.00
5 5.16 94.83
Exchange Rate 10 5.16 94.83
Lags: Nifty Export
2 99.85 0.14
5 99.46 0.53
Nifty 10 99.45 0.54
2 2.43 97.56
5 8.54 91.45
Export 10 8.63 91.36
Lags: Nifty IIP
2 98.31 1.68
5 98.35 1.64
Nifty 10 98.35 1.64
2 10.79 89.20
5 12.59 87.40
IIP 10 12.63 87.36
Lags: Nifty Interest Rate
2 99.99 0.01
5 92.94 7.059
Nifty 10 92.67 7.32
2 0.23 99.76
5 3.49 96.50
Interest Rate 10 3.67 96.32
Lags: Nifty Money Supply
2 99.43 0.56
5 99.43 0.56
Nifty 10 99.43 0.56
2 0.12 99.87
5 0.12 99.87
Money Supply 10 0.12 99.87
Lags: Nifty FDI
2 96.46 3.53
5 95.77 4.22
Nifty 10 95.70 4.29
2 0.65 99.34
5 1.91 98.08
FDI 10 1.92 98.08
Lags: Nifty Sensex
2 9.98 90.01
5 9.71 90.28
Nifty 10 9.69 90.30
2 3.21 96.78
5 3.10 96.89
Sensex 10 3.10 96.89

Similarly, the variance decomposition results at the end of 10 periods for BSE Sensex in
bivariate framework are shown in Table 8. In case of bivariate modeling of BSE Sensex and exchange
rate, BSE Sensex explains nearly 98% of its own forecast error variance while Exchange rate explains
only 2 percent of BSE Sensex. Results indicate Exchange Rate explains over 91% of its own variance
and Sensex explains remaining 9 percent of variance of exchange rate. It implies that BSE Sensex

Electronic copy available at: https://ssrn.com/abstract=1693544

160 International Research Journal of Finance and Economics - Issue 14 (2008)

causes Exchange Rate. Results indicate that BSE Sensex explains almost 99 percent of its variance and
about 3 percent of variance of exports while export explains about 1 percent of variance in BSE Sensex
and 97 percent its own variance. BSE Sensex seems to cause export flows. In case of Sensex and IIP
BVAR, Sensex explains 5 percent of variance of IIP while IIP explains about 1 percent of variance in
BSE Sensex. In this case, Sensex causes IIP. Sensex-Interest BVAR Rate results indicate that interest
rate explains 6.41 percent variance in BSE Sensex while BSE Sensex explains only 0.12 percent of
Interest rate, thus Interest Rate causes BSE Sensex. Money supply only explains 2.44 percent of BSE
Sensex while 6.90 percent of variance of Money supply is explained by BSE Sensex in case of Sensex-
Money Supply BVAR model. Direction of effect seems from BSE Sensex to Money Supply. It is
difficult to infer in case of Sensex-FDI BVAR, as small proportion of variance of each other are
explained in this case. FDI explains approximately 3 percent of Sensex variance while BSE sensex
explains 2 percent of FDI variance. However, FDI may be causing movement in BSE Sensex in this
case.

Table 8: Variance Decomposition of Sensex and other variables

Variance Decomposition of At Time Period Sensex Exchange Rate

2 98.71 1.28
5 98.40 1.59
Sensex 10 98.40 1.59
2 8.37 91.62
5 9.42 90.57
Exchange Rate 10 9.42 90.57
Sensex Export
2 99.78 0.21
5 99.04 0.95
Sensex 10 99.04 0.95
2 0.88 99.11
5 2.88 97.11
Export 10 3.08 96.91
Sensex IIP
2 99.19 0.80
5 98.45 1.54
Sensex 10 98.37 1.62
2 2.14 97.85
5 5.08 94.91
IIP 10 5.06 94.93
Sensex Interest Rate
2 94.85 5.14
5 92.53 7.46
Sensex 10 92.25 7.74
2 1.43 98.56
5 3.46 96.53
Interest Rate 10 3.80 96.19
Sensex Money Supply
2 99.65 0.34
5 99.65 0.34
Sensex 10 99.65 0.34
2 1.65 98.34
5 1.65 98.34
Money Supply 10 1.65 98.34
Lags: Sensex FDI
2 99.95 0.04
5 97.14 2.85
Sensex 10 96.97 3.02
2 1.16 98.83
5 1.95 98.04
FDI 10 2.09 97.90

To investigate dynamic responses further between the variables, Impulse Response of the VAR
system has also been estimated. An Impulse Response function traces the effect of a one-time shock to

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 161

one of the innovations on current and future values of the endogenous variables. So, for each variable
from each equation separately, a unit shock is applied to the error, and the effects upon the VAR
system over time is noted. A shock to the i-th variable not only directly affects the i-th variable but is
also transmitted to all of the other endogenous variables through the dynamic (lag) structure of the
VAR. Thus, if there are g variables in a system, total of g2 impulse responses could be generated. In
our case, we have thirteen BVAR pairs of variables; technically four graphs of Impulse Responses (IR)
for each BVAR pair are possible in each period totaling fifty two. However, we have presented only
twenty-six IR graph for the sake of brevity because IR graph representing direct affects on the i-th
variable have not been presented as it does not reveal any causal pattern of response in theses cases. As
indicated by variance decomposition, similar pattern of causality is also observed graphically using
impulse response functions. Further, Impulse Response Function indicates that the BVAR system is
stable.

6. Concluding Remarks
The purpose of the present study is to explore the causal relationships between stock prices and the key
macro variables representing real and financial sector of the Indian economy. These variables are the
index of industrial production, exports, foreign direct investment, money supply, exchange rate,
interest rate, NSE Nifty and BSE Sensex. The present analysis is based on quarterly data from March,
1995 to March, 2007. Johansen’s approach of cointegration and T-Y Granger causality test have been
applied to explore the long-run causal relationships while BVAR modeling for variance decomposition
and impulse response functions has been applied to examine short run causal relationships.
Co-integration regressions indicate the presence of a long run relationship between stock prices
and FDI, stock prices and MS and stock prices and IIP. However, except for interest rate and exports,
the pattern of cointegration and long-run Granger causality in each market indicates differential
pattern. In NSE, movement in NSE does not have effect on exchange rate and IIP while movement in
BSE Sensex seems to cause these variables. In this context, it is important to highlight that the
composition of NSE Nifty and BSE Sensex vary from each other. Nifty is a benchmark index that is
composed of 50 stocks i.e. shares of 50 companies. Sensex, on the other hand, is an index that is
composed of thirty stocks i.e. shares of 30 companies. The rate of change in variation may happen
because every stock in an index has some index-weightage and it is varying in each index. In case of
short run, there is no differential impact of causal pattern as indicated by variance decomposition and
impulse response functions. In this case, the results reveal that NSE Nifty causes exchange rate,
exports, IIP and money supply while interest rate and FDI causes NSE Nifty. Broadly, these patterns
are valid also in the case of BSE Sensex.
The study reveals that movement in stock prices causes movement in IIP. It clearly implies that
stock prices leads real economic activity. It reveals that growth rate of real sector is factored in the
movement in stock prices. One possible explanation is that major players such as FIIs and mutual
funds are well aware in advance of business plan of major industrial houses. Another explanation is the
higher demand of stocks created by positive growth expectations. Higher demand has been created by
increased due to increased participation of domestic retail investors and trader directly or through
mutual fund as the expected returns in alternative investment opportunities such as bank deposits,
deposits in post offices, etc are less than the expected return in stocks. Further, FIIs inflows has
increased in recent years due to expected higher growth of rate of return in Indian compared to USA,
EU and other developed economies. This has been possible due to change in legal and institutional
trading environment. This study also reveals that FDI does cause movement in stock prices while
movement in stock prices does not have significant effect on FDI. This links may be the outcome of
positive effect of FDI inflows on expectations regarding growth, earning and profit prospect of the
sector. Results indicate that the movements in stock prices seem to affect export flows, possibly
through its effect on exchange rate. It is to be noted that the movement in Sensex and Nifty are causing
changes in exchange rate at least in the short run. One possible inference may be drawn from the

Electronic copy available at: https://ssrn.com/abstract=1693544

162 International Research Journal of Finance and Economics - Issue 14 (2008)

present study is that the stock market movement in India seems to be driven by the performance of
industry in domestic market rather than in export market. It seems to be intuitively correct as the export
segment of major listed companies is small barring IT and textile sector.
The study indicates that exchange rate does not influence the BSE Sensex and NSE Nifty while
movement in BSE Sensex and NSE Nifty does cause change in exchange rate. It indicates towards
large FII inflows in recent years. It is causing serious problem to the real sector of the economy. As of
October, the rupee had appreciated close to 13 per cent on a year-on-year basis against the greenback.
Nearly 7.5 per cent of this appreciation came in the months of April and May. In recent months, net FII
inflows have picked up — around $5.7 billion in October itself. The recent cuts in the Fed rate are
likely to provide a further impetus to the inflows and put more pressure on the rupee. One of the major
worries of rupee appreciation is the hit exports are expected to take.
The relationship between money supply and stock prices has been widely studied because of
the belief that money supply changes have important direct effects through portfolio changes, and
indirect effects through their effect on real economic activity, which in turn postulated to be the
fundamental determinants of stock prices. However, the study indicates that money supply does not
affect movement in stock market indices while interest rate causes change in stock indices. The
possible explanation may be the lack of understanding on the part of players in the stock market about
the complex linkage of money supply with cost and profit margin of industry. Further, major player are
either not able to predict the possible effect of money supply and monetary measures related to money
supply taken by RBI or they could not form their expectation on the basis of changes in money supply.
Given the demand for money, there is a direct inverse relationship between money supply and interest
rate. Probably, stock market reacts to money supply only if it causes changes in interest rate. It may be
inferred that monetary policy measures affecting interest rate are critical in influencing stock market in
India.

Electronic copy available at: https://ssrn.com/abstract=1693544

International Research Journal of Finance and Economics - Issue 14 (2008) 163

References
[1] Abdalla, I. S. A. and V. Murinde (1997), “Exchange Rate and Stock Price Interactions in
Emerging Financial Markets: Evidence on India, Korea, Pakistan, and Philippines,” Applied
Financial Economics 7, 25-35.
[2] Abdalla, I. S. A. and V. Murinde (1996) Exchange rate and stock prices interactions in
emerging financial markets: Evidence on India, Korea, Pakistan and Philippines. Applied
Financial Economics, 7, 25-35.
[3] Ando, A. and F. Modigliani (1963), “The Life Cycle Hypothesis of Saving: Aggregate
Implications and tests,” American Economic Review, Vol. 53, No. 1.
[4] Ajayi, R. A. and M. Mougoue (1996), “On the Dynamic Relation between Stock Prices and
Exchange Rates,” Journal of Financial Research 19, 193-207.
[5] Bahmani-Oskooee, M. and A. Sohrabian (1992). “Stock Prices and the Effective Exchange
Rate of the Dollar”, Applied Economics, 24, 4, 459-464.
[6] Baumol, W. (1965), Stock Market and Economic Efficiency, Fordham University Press, New
York.
[7] Bernanke, B.S. and K.N. Kuttner (2005), “What Explains the Stock Market’s Reaction to
Federal Reserve Policy,”Journal of finance, 60, 1221-1257.
[8] Bewley, R., and Orden, D.(1994), “Alternative methods of estimating long run responses with
applications to Australian import demand”, Econometric Reviews, 13, 179-204.
[9] Blough, S. R. (1992), “The Relationship Between Power and Level for Generic Unit Root Tests
in Finite Samples”, Journal of Applied Econometrics, 7, 295–308.
[10] Bhattacharya B and Mukherjee J. (2002), “Causal relationship between stock market and
exchange rate, foreign exchange reserves and value of trade balance: A case study for India”,.
www.igidr.ac.in
[11] Bodie, Zvi. (1976), “Common Stocks As A Hedge Against Inflation,” Journal of Finance,
31(2), 459-470.
[12] Bosworth, B.(1975), "The Stock Market and the Economy", Brookings Papers on Economic
Activity, Vol. 2.
[13] Chen, N, R. Roll, and S. Ross (1986), “Economic forces and the stock market”, Journal of
Business 59, 383-403.
[14] Chou, Y L (1969): Statistical Analysts, Holt Rinehart and Winston, London.
[15] Cutler, D. M., J. M. Poterba, and L. H. Summers (1989), “What moves stock prices?”, Journal
of Portfolio Management, 15, 4-12.
[16] Dickey, D. A. and W. A. Fuller (1979), “Distribution of the Estimators for Autoregressive Time
Series with a Unit Root,” Journal of the American Statistical Association 74, 427-431.
[17] Dickey, D. A. and W. A. Fuller (1981), “Likelihood Ratio Statistics for Autoregressive Time
Series with a Unit Root,” Econometrica 49, 1057-1072.
[18] Dolado, J.J. and Lutkepohl, H. (1996), “Making Wald Test Work for Cointegrated VAR
Systems”, Econometrics Reviews, 15, 369-386.
[19] Dornbusch, R. and S. Fischer (1980), “Exchange Rates and Current Account”, American
Economic Review, 70, 960-971.
[20] Enders, W. (1995), Applied Economic Time Series, Wiley, New York.
[21] Engle, R. F. and C. W. J. Granger (1987), “Co-integration and error-correction: Representation,
estimation and testing”, Econometrica, 55, 251—276.
[22] Fama, E. F. (1981), “Stock Returns, Real Activity, Inflation, and Money”, American Economic
Review, 71(4), 545-565.
[23] Friedman, M. (1988), “Money and the Stock Market”, Journal of Political Economy, 96, 2,
221-245.
[24] Geske, R. and R. Roll (1983), “The Fiscal and Monetary Linkage between Stock Returns and
Inflation,” Journal of Finance, 38(1), 1-34.

Electronic copy available at: https://ssrn.com/abstract=1693544

164 International Research Journal of Finance and Economics - Issue 14 (2008)

[25] Giles, J.A. and Mirza, S. (1999), “Some pretesting issues on testing for Granger noncausality”,
Econometrics Working Paper EWP 9914, Department of Economics, University of Victoria.
[26] Granger, C. W. J. (1988), “Some Recent Developments in the Concept of Causality,” Journal of
Econometrics, 39, 199-211.
[27] Granger, C. W. J. and Newbold, P. (1974), “Spurious Regression in Econometrics” Journal of
Econometrics, 2, 111-120.
[28] Granger, C. W. J. and Newbold, P. (1974), “Spurious Regression in Econometrics” Journal of
Econometrics, 2, 111-120.
[29] Gujarati, D.N.(2003), Basic Econometrics, McGraw Hill, India.
[30] Ho, M. S., and B. E. Sorensen (1996), “Finding cointegration rank in high dimension systems
using the Johansen test. An illustration uses data based Monte Carlo simulations”, Review of
Economics and Statistics, 78, 726- 732.
[31] Johansen, S. (1991), “Estimation and hypothesis testing of cointegration vector in Gaussian
vector autoregressive models”, Econometrica, 59, 1551-1581.
[32] Kwiatkowski, D., Philips, P.C.B., Schmidt, P., Shin, Y., (1992), “Testing the Null Hypothesis
of Stationary against the Alternative of a Unit Root. How Sure Are We that Economic Time
Series Have a Unit Root?”, Journal of Econometrics, 54, pp.159-178.
[33] Ma, C. K. and G. W. Kao (1990), “On Exchange Rate Changes and Stock Price Reactions,”
Journal of Business Finance and Accounting, 17, 441-449.
[34] Maddala G.S. and I.M. Kim (1998), Unit Roots, Cointegration and Structural Change,
Cambridge University Press.
[35] Mukherjee, T.K. and A. Naka, (1995), “Dynamic relations between macroeconomic variables
and the Japanese stock market: An application of a vector error correction model”, The Journal
of Financial Research, 2, 223-237.
[36] Pan, Fok & Lui (1999), Dynamic Linkages Between Exchange Rates and Stock Prices:
Evidence from Pacific Rim Countries, Shippensburg University Working Papers.
[37] Parkinson, J M (1987): ‘The EMH and the CAPM on the Nairobi Stock Exchange”, East
African Economic Review, 13, 105-110.
[38] Pearce, D. K. and V. V. Roley (1983), “The Reaction of Stock Prices to Unanticipated Changes
in Money: A Note”, Journal of Finance 38, 1323-1333.
[39] Pearce, D.K. (1985), “Stock Prices and Economic News”, Journal of Business, 58(1), 49-67.
[40] Phillips, P. C. and Perron, P. (1988), “Testing for a Unit Root in Time Series Regression”
Biometrika, 75, 335-346.
[41] Sims, C. A. (1980), ‘Macroeconomics and Reality’, Econometrica, 48(1): 1-48.
[42] Toda, H.Y. and T. Yamamoto (1995), “Statistical inferences in vector autoregressions with
possibly integrated processes”, Journal of Econometrics, 66, 225-50.

Electronic copy available at: https://ssrn.com/abstract=1693544