An Empirical Analysis of Stock Market Performance and Economic Growth: Evidence from India

Unit Root Tests

The study starts with the conventional unit root tests, to find out the order of integration by using the Augmented Dickey-Fuller (Dickey and Fuller, 1979) test, Phillips-Perron (Phillips and Perron, 1988) test and the KPSS (Kwiatkowski et al. 1992) test. All of these unit root tests are used to test whether the data contains unit root (non-stationary) or is a stationary process. A series is said to be stationary if the mean and auto co variances of the series do not depend on the time factor. Any series that is not stationary then it is said to be non-stationary. A series is said to be integrated of order ‘d’ which can be denoted by I (d), means that it has to be differenced ‘d’ times before it becomes stationary. Otherwise, if a series by itself, let say stationary at levels, without having to be differenced, then that is said to be I (0). It is very essential to apply unit root tests for individual series to come up with some idea that whether the variables are integrated with same order or not, if the order of integration is same for the entire variables then it is quite possible that study can find out the long run and short run dynamic behavior of the variables by employing Engle-Granger cointegration test and error correction model. Hence, if the order of integration is not same for all the variables then it’s not possible to use above methodology (cointegration and error correction model) for the analysis. In case of both ADF and PP
tests, the null hypothesis of non-stationary (unit root) is tested against the alternative hypothesis of stationarity.

For the Augmented Dickey-Fuller (ADF) test; consider a simple AR (1) process:

y_t  
t

t


^y t 1 ^

x ^'  

(3.1)

Where

y_t is the observed variable and

x_t are optional exogenous regressors which may consist

of constant or a constant and trend,  and  are parameters to be estimated, and the  _t

are assumed to

be white noise (i.e., zero mean and constant variance). _If 

^{ 1}, ^y is a non-stationary series and the

variance of ^y increases with time and approaches infinity, on the other hand if ^{  1}, then ^y is a

^{(trend) stationary series. Now, subtracting equation (1) both sides with y}t 1 ^{, then we get:}
x



'

 y _t

  y

t  1 ^ t

• 
   _t
  

(3.2)

Where     1. The null and alternative hypotheses can be written as;

H ₀ :

H₁ :

 = 0 ( ^t

  0 ( ^t

is unit root) is stationary)

and can be evaluated using the conventional t -ratio for :