Page 48 - IILMGSM Journal_Management Perspective
P. 48
st of all for analyzing a time series it is important X if the Ö ’s, k>0, coefficients of Y are individually
to check for stationary properties. Granger’s
causality test can be applied for a stationary time tk t
series only. A time series is said to be (weakly)
stationary if its mean and variance are constant over and / or jointly different from zero.
time and the value of the covariance between the two
time periods depends only on the distance or gap or To examine whether there is any long run equilibrium
lag between the two time periods and not the actual relationship among the segments considered
time at which the covariance is computed [Gujrati Johansen test is applied. It is likelihood ratio test
and Sangeetha (2007)] for the co integration and specifies the number of co
integrating relationships. Multivariate and bivariate
co integration test has been conducted.
Empirical Results
To check whether the index series is stationary or As mentioned earlier the whole time period of the
not unit root tests - Augmented Dickey Fuller (ADF) study (1994 to 2008) has been divided into three
Test and Phillip Perron (PP) Test are applied. All sub periods:
the series are found to be integrated of order 1 [I
(1)]. 1) January 1994- May 2000
A time series X Granger-causes another time series 2) June 2000 to December 2004
t
3) January 2005 to December 2008
Y if the later can be predicted with better accuracy
t The analysis has been conducted on the whole
time period and the sub periods.
by using past values of X rather than by not doing
t Table 2 summarizes the descriptive statistics of the
variables for all sub periods. Looking at the mean
so, other information being identical. Testing causal values it is observed that PLR and DR stay around
the same levels with a slight increase in the last sub
relations between two stationary series ÄX and ÄY period. In the money market the mean interest rates
tt fall sharply in the second sub period and there is
slight increase in the third sub period. Therefore, the
can be based on the following two equations: third sub period is the period of rising interest rates.
NIFTY is found to be increasing continuously across
pp the sub periods. Mean spot exchange rate is found
to increase from first sub period to second. In the
Yt 0 k Ytk k X tk t third sub period it falls to Rs.43.56 per Dollar.
k 1 k 1
pp
X t 0 k X tk Ytk t
k 1 k 1
Where Ä is the difference operator, Y t-k and Xt-k
represent the lagged value of Y and X , ì and õ are
t tt t
disturbance terms assumed to be white noise. The
lag length (k = 1, 2, …., p) have been chosen
arbitrarily with maximum of six lags. The null
hypothesis that X does not Granger-cause Y is not
tt
accepted if the â ’s (k>0) are individually as well as
k
jointly significantly different from zero using
standard test (F test). Similarly, Y Granger-causes
t
(47)
to check for stationary properties. Granger’s
causality test can be applied for a stationary time tk t
series only. A time series is said to be (weakly)
stationary if its mean and variance are constant over and / or jointly different from zero.
time and the value of the covariance between the two
time periods depends only on the distance or gap or To examine whether there is any long run equilibrium
lag between the two time periods and not the actual relationship among the segments considered
time at which the covariance is computed [Gujrati Johansen test is applied. It is likelihood ratio test
and Sangeetha (2007)] for the co integration and specifies the number of co
integrating relationships. Multivariate and bivariate
co integration test has been conducted.
Empirical Results
To check whether the index series is stationary or As mentioned earlier the whole time period of the
not unit root tests - Augmented Dickey Fuller (ADF) study (1994 to 2008) has been divided into three
Test and Phillip Perron (PP) Test are applied. All sub periods:
the series are found to be integrated of order 1 [I
(1)]. 1) January 1994- May 2000
A time series X Granger-causes another time series 2) June 2000 to December 2004
t
3) January 2005 to December 2008
Y if the later can be predicted with better accuracy
t The analysis has been conducted on the whole
time period and the sub periods.
by using past values of X rather than by not doing
t Table 2 summarizes the descriptive statistics of the
variables for all sub periods. Looking at the mean
so, other information being identical. Testing causal values it is observed that PLR and DR stay around
the same levels with a slight increase in the last sub
relations between two stationary series ÄX and ÄY period. In the money market the mean interest rates
tt fall sharply in the second sub period and there is
slight increase in the third sub period. Therefore, the
can be based on the following two equations: third sub period is the period of rising interest rates.
NIFTY is found to be increasing continuously across
pp the sub periods. Mean spot exchange rate is found
to increase from first sub period to second. In the
Yt 0 k Ytk k X tk t third sub period it falls to Rs.43.56 per Dollar.
k 1 k 1
pp
X t 0 k X tk Ytk t
k 1 k 1
Where Ä is the difference operator, Y t-k and Xt-k
represent the lagged value of Y and X , ì and õ are
t tt t
disturbance terms assumed to be white noise. The
lag length (k = 1, 2, …., p) have been chosen
arbitrarily with maximum of six lags. The null
hypothesis that X does not Granger-cause Y is not
tt
accepted if the â ’s (k>0) are individually as well as
k
jointly significantly different from zero using
standard test (F test). Similarly, Y Granger-causes
t
(47)
