EViews.com

Hi,

I am trying to write a EViews code for a simple Black Litterman model. Since i am a relatively new to eviews world, just wanted to enquire

---Do we have any built in command for solving Black Litterman with input parameters?

If not then

--What is the command to solve a quadratic program in EViews?

Thanks!


Cheers,
Gaurav

gs2440 wrote:Do we have any built in command for solving Black Litterman with input parameters?

Please read this (i.e. the second post).

gs2440 wrote:What is the command to solve a quadratic program in EViews?

You can either use the LogL object or try the matrix language.

Hi there,

i deal with the same problem at the moment and can't find any help/approach in the www.

Isn't there anyone who has come to this problem before and may can give some initial advice?

How can i solve in my program the minimum-variance-portfolio for 2 assets with constraints?
In Matlab e.g. i would run the "quadrativ programming solver" with the Covarianve-Matrix as Input.
And i would set the constraints as no short selling ( solution-vector x_i > 0 for all i) and the sum of the weights should equal 1 ( sum of x_i = 1).
Additionaly i would set my target value in the portfoliomean.

Like here: http://www.styleadvisor.com/support/sta ... iance.html

Regards, Benjamin

You can try EViews 7, since this new version can interact with MATLAB. In the meantime, the code below might be of some help. Open a LogL object (Object/New Object/LogL) and paste the following:



The specification above is one of the many ways to build such a model. And different modifications (including the initial values) might yield better results depending on your data and the type of your problem.

You should change the name of series (i.e. x1 and x2) to match yours as well as the desired target level of portfolio mean. Please keep in mind that you should compute the @logit(c(1)) to obtain the actual weight. You may experience convergence difficulties, since LogL object is not really meant for general purpose constrained optimization.