Help deriving second order ECM please

[sarah00]

Hi,
I need some help deriving a second order ECM from a second order ARDL please.

I know how to do it from a first order ARDL, but am getting confused with a higher order. I know the end result of the ECM2 but having problems getting there.

This is the procedure I've been using:
1. Starting from ARDL(2) yt = a0 + a1Xt +a2Xt-1 + a3Xt-2 + b1yt-1 + b2yt-2 + et minus yt-1 from both sides

2. Then add and subtract a1Xt-1 to right hand side.

3. Add and subtract a2Xt-2 on RHS

4. Add and subtract b1yt-2 on RHS

Is this correct? I always seem to end up with a second lag for x and y, but for ECM2 the maximum lag is one.

Can anyone help please?

[sarah00]

Am I making a really obvious mistake? Like because of co-linearity the second lag cancels? I really need some help to be sure