seen that when yield is a dependent variable for BSE SENSEX, the coefficient of P/E ratio for lag (–4)
is negative and only significant coefficient (at 10 per cent significant level). The results indicate that
if the P/E goes up, then in 4 months’ time yield will go down. A Granger causality test indicates that
the P/E ratio Granger-causes yields to change. It can be seen from the table that chi-square test for
pair-wise Granger causality rejects the null hypotheses (at 5 per cent significant level) that P/E ratio
does not Granger-cause yield.
When VAR is conducted for BSE SENSEX where price is taken as a dependent variable, the coefficient
of P/E ratio for lag (–4) is found to be positive and significant (at 5 per cent significant level) coefficient.
The results confirm that if the P/E ratio goes up, then in 4 months’ time price will go up. A Granger
causality test indicates that the P/E ratio Granger-causes prices to change for BSE SENSEX. It can be
seen from the Table that chi-square test for pair-wise Granger causality rejects the null hypotheses (at
5 per cent significant level) that P/E ratio does not Granger-cause price.
Tables 5B–5D show results for the VAR analysis for P/E ratios and prices/yields. Since no cointegration
was found for other indexes, error correction was not required. For S&P BSE 500 when price is taken as
a dependent variable, the lagged coefficient of P/E (–3) is found to be negative and significant (at 5 per cent
significant level) indicating that increase in P/E leads to decrease in price for S&P BSE 500. The relation
between yield and lagged coefficient P/E (–4) is also negative and significant (at 5 per cent significant
level) indicating decrease in yield with the increase in P/E. Granger causality test indicates that the P/E
ratio Granger-causes price change for S&P BSE500. However, chi-square with yield as a dependent
series accepts the null hypothesis that P/E does not Granger-cause yield.