[R-SIG-Finance] constrained weighted least square estimation

[Philipp]

Dear all,

First of all I am not proffessional in R and sorry if my question is not
very good formulated.
So I have a problem I want to estimate a constrained regression with a
weighted least square estimator (Rouwenhorst modell 1994). In fact it is a
regression with dependent variable as index return and the independent
variables are Dummyvariables. 

index_returns(ij)=a + sum(beta(i)*I(i))*w(i))+sum(gama(j)*C(j)*v(j))  +
e(ij)

with index_returns in country j and Industry i (I have Industry indexes
country-specific, that means I have more observations than parameters to
estimate) and a=global return same for all industries and countries,
beta(i)=Coefficient for Industry I(i). I(i) is one if index return is from
industry i. C(j) is one if index return is from country j. Gamma(j) the
Coefficient for Country j. w(i) is the market value of industry i divided by
the total market capitalization in time t.
v(i) the market value of country j divided by the total market
capitalization in time t. 
In order to avoid multicolinearity two linear constraints are integrated:

sum(w(i)*beta(i)=0
sum(v(j)*gamma(j)=0

So this is my problem. A Cross sectional constrained weighted least square
regression. I tried it for example with Quadratic Programming. But I am not
sure how to integrate constraints and by the way I need standard errors. 
This for my master thesis and since everything i wanted to analyse is build
up on this Regression, i am quite stucked, since i tried everything.

Could Somebody help me

Kind regards

Phillip




--
View this message in context: http://r.789695.n4.nabble.com/constrained-weighted-least-square-estimation-tp4205373p4205373.html
Sent from the Rmetrics mailing list archive at Nabble.com.