interpretation of long run coefficient - VECM

[NIck Stavr]

Hello, I had issues pasting the table in this box so I posted it below.
I am using a VECM and I read that I should reverse the sign of the coefficient of my variable to interpret the actual relationship but I am a bit unsure if someone could clarify this please. For example the interpretation of the table above is :
1) lnPOIL has a negative and statistical significant impact on lnGDP
2) lnREC has a positive and statistical significant impact on lnGDP
Are these two statements correct?
Thank you in advance

[NIck Stavr]

This is the table I am referring to, I try several ways to paste it and readjust the coefficients but as soon as I post it it all goes back to being squeezed. I hope you can understand what is going on. I am interested in the long-run causality of lnPOIL and lnREC. Are my interpretations above correct? Again thank you

	Johansen	normalization	restriction imposed
beta	Coef.	Std. Err.	z P>z [95% Conf.	Interval]
_ce1
lnGDP	1	.	. . .	.
lnCPI	-.140817	.1361749	-1.03 0.301 -.4077148	.1260808
lnPOIL	.4479317	.1995608	2.24 0.025 .0567996	.8390637
lnREC	-2.015953	.6516066	-3.09 0.002 -3.293078	-.7388276
_cons	3.378265	.	. . .	.

[Sebastian Kripfganz]

Yes, you could think about this as the following equilibrium relationship:

lnGDP -0.140817 lnCPI +0.4479317 lnPOIL -2.015953 lnREC +3.378265 = 0

Once you write lnGDP as a function of the other variables by moving them to the right-hand side of the equation, you see why you need to reverse the sign.

[NIck Stavr]

Hello Sebastian, thank you for your reply.
I have seen somewhere online that the equation is like this: lnGDP -0.140817 lnCPI +0.4479317 lnPOIL -2.015953 lnREC +3.378265 = ECT
is this different ? or are they approximately the same?
thank you again

[Mead Over]

If by ECT is meant Epsilon with subscripts c and t (perhaps for country and year???), and if you assume that in equilibrium the expected value of Epsilon_c_t is zero, then your version of the equilibrium condition is just the stochastic version of the expression in post #3.

[NIck Stavr]

no the ECT is the error correction term which is added in the VECM, however your explanation in #5 may still be valid but I am not sure, that's why I asked for a confirmation.
Thank you

[Sebastian Kripfganz]

What I have written in my previous post with 0 on the right-hand side was the equilibrium relationship. The error correction term (ECT) is a deviation from this equilibrium, when the right-hand side is possibly nonzero. (In practice, the variables will almost always be out of equilibrium. The ECT is then the stochastic disturbance to that equilibrium, in line with what Mead said.)