<t>Assume we have two portfolios: W and IW has stocks A,B,C with w1,w2,w3 weightsI has stocks A,B with i1,i2 weightHow to modify existent stock correlation matrix in order corr(I,W) would increase more, then corr(A,W) and corr(B,W)?I though modifying stock corr matrix as: alpha*oldCorr+(1-alpha)*1 w...

did you manage to get an answer? i would be interested to know too.thanks

<t>Please help!!!!!!When we are building a Hull trinomial tree, should we treat sigma as volatility of changes in rates or vol of returns (using log differences) or volatility of rates?I would think that we suppose to use vols of differences, right?And assuming that i know vols of returns can i conv...

right, i am interested in 3 factors in price movements.how about a second approach which i described below (through calculation principal components of price movements), is this method valid?thanks

<t>no options .....so I will try to reformulate my question:A corr matrix A=V*D*V', where V-eigenvector, D- egenvaluesI decided to use 3pcs, so I set all eigenvalues to 0 except first 3D_new=[D(1,1) D(2,2) D(3,3) 0 0 0 ....]And now I want to caclulate corr matrix with only 3pcs.But, it doesn't look ...

<t>thanks for your answers. But there is another question:I did PCA on correlation matrixA=VDV'where V eigenvectors, D eigenvalues then I decided to use only 3PCs, which explained 97% of the total variance and calculated new correlation matrix using only 3PCs:A*=VDV'as a result on the main diagonal ...

can HW model be used to price long term derivatives, say with duration of 30yr?

sorry for stupid question. but i am confused.I calculated eigenvalues and eigenvectors on corr matrix and now i want to restore corr matrix using only first 3 pca. how to do it?thanks.

thanks Aaron.i checked for outliers, and tried to remove them, but it cause only mor negative correlation (probably it was not outliers at all, but real jumps in interest rates).can you advise easy methods to smooth data before calculate correlation?thanks

thanks Aaron.i checked for outliers, and tried to remove them, but it cause only mor negative correlation (probably it was not outliers at all, but real jumps in interest rates).can you advise easy methods to smooth data before calculate correlation?thanks

I've got a negative correlation between forward rates differences (calculated from libor fixings) using 10 yr history.ex, corr(diff(forward(30,60)),diff(forward(60,90))) =-0.32 is it possible? thanks