piecewise regression in Eviews
Posted: Mon Feb 28, 2011 12:42 pm
Hi all,
I have consulted this forum's search function, google and quite some litterature, however I am yet to find a solution.
I would like to perform a piecewise linear regression as follows:
Rp - Rf = (1 - D)[Aml + Bml(Rm - Rf)] + (D)[Amh + Bmh * (Rm - Rf)]
Where
Rp = portfolio returns
Rf = risk-free rate
Rm = market return
A = alpha
B = Beta
D = treshold dummy (equal to one if market return greater than treshold, zero otherwise)
suffixes ml and mh denotes market high or market low, respectively
Also I would like a continuous line, so a restriction like Aml + Bml(Treshold) = Amh + Bmh(reshold) would probably be suitable
Can Eviews do this, and if so, what would I need to do in order to get it to comply?
I know a bit about EViews but almost NONE about programming it, anyone that would like to help, anything would be much appreaciated!
Rgds
I have consulted this forum's search function, google and quite some litterature, however I am yet to find a solution.
I would like to perform a piecewise linear regression as follows:
Rp - Rf = (1 - D)[Aml + Bml(Rm - Rf)] + (D)[Amh + Bmh * (Rm - Rf)]
Where
Rp = portfolio returns
Rf = risk-free rate
Rm = market return
A = alpha
B = Beta
D = treshold dummy (equal to one if market return greater than treshold, zero otherwise)
suffixes ml and mh denotes market high or market low, respectively
Also I would like a continuous line, so a restriction like Aml + Bml(Treshold) = Amh + Bmh(reshold) would probably be suitable
Can Eviews do this, and if so, what would I need to do in order to get it to comply?
I know a bit about EViews but almost NONE about programming it, anyone that would like to help, anything would be much appreaciated!
Rgds