cointegration

[nayan]

Dear sir,

i am currently using Johansten Mutivariate cointegration to examine the interaction between macroeconomic factors and stock market. I found the cointegration but I am confused with the Error correction model (ECM). There are two variants of error correction model. I dont know what is the difference betwen two?

The first eqn is what eviews use for ECM (eqn 24.23 eviews 5 users guide )
The second variant uses t-p in place of t-1 for subscript of y.

Moreover, the ECM used in most of the empirical study uses the summation from 1 to p (lag) rather than p-1.Is this the ECM of granger causality . If so what is the difference between this and Vector error correction model of Johansten cointegration?

Thanking you