Dear,

For our master thesis we need to regress the equation
Pt-P(t-1)= S/2*(Qt-Q(t-1))+Beta*S/2*(Qt-1)+et

or http://prntscr.com/6q81o8

It is the model of Huang Stoll. The paper can be found here: http://www.acsu.buffalo.edu/~keechun...k%20spread.pdf

Our question is how can we model the first variable term in the regression in eviews or matlab:S/2*(Qt-Q(t-1)). We have subtracted it from Pt-P(t-1) in excel and used as depentent variable in eviews, but we don't know if that is allowed. The hardest part is that we have variables S, Qt and Q(t-1) that have influence on the dependent variable but do not have a coefficient. So we calculated S/2*(Qt-Q(t-1)) and S/2*(Qt-1) in excel. Does anybody has more information on this kind of regression or knows how this is called.

Thank you so much!