EViews.com

Good morning,
given the versatility of OLS models with arbitrary inequality constraints, I was wondering whether it would be possible to implement OLS with inequality constraints, for example using the approach laid out here (http://www.jstor.org/discover/10.2307/2 ... 2140463023), in Eviews. I am aware of the workarounds to impose simple constraints via monotonic transformations, but "real" inequality constraints would make life a whole lot easier. The workarounds also do not work with limited data availability, e.g. 25 observations are insufficient to use a transformation that imposes 0<ci<1 and c1+c2+c3...=1.
Given my limited programming experience, and my full-time job in a different function at IHS, I currently lack the necessary skill and time to undertake this venture myself.

I'm not sure what you're asking here...

I was wondering whether somebody has already written a program that can handle OLS with inequality constraints, or whether such a feature is planned for Eviews in the near future. Matlab, for instance, has the function "lsqlin" which efficiently estimates OLS with inequality constraints and equality constraints simultaneously. In practical work, this feature would really make life a lot easier.

I'm aware of no such program.

I was wondering whether somebody has already written a program that can handle OLS with inequality constraints, or whether such a feature is planned for Eviews in the near future. Matlab, for instance, has the function "lsqlin" which efficiently estimates OLS with inequality constraints and equality constraints simultaneously. In practical work, this feature would really make life a lot easier.

It might be easiest to call lsqlin directly by connecting EViews to Matlab. (I imagine there is an equivalent R routine as an alternative.)