Serving the Quantitative Finance Community

 
User avatar
player
Topic Author
Posts: 0
Joined: August 5th, 2002, 10:00 am

optimized basket

April 2nd, 2010, 11:00 am

A trivial question but does anyone know how i would start to test an optimized basket against an index but also taking into account bid/ask spreads on the underlying instruments?
 
User avatar
Errrb
Posts: 0
Joined: December 17th, 2002, 4:18 pm

optimized basket

April 3rd, 2010, 2:32 pm

I assume you want to estimate the tracking error between your basket and the index. In this case you can compute bid/ask value of you basket and compare it with bid/ask of the complete index, the part of tracking error will be due to spreads then as you expected. Not sure I answered the question which you had in mind.
 
User avatar
player
Topic Author
Posts: 0
Joined: August 5th, 2002, 10:00 am

optimized basket

April 3rd, 2010, 5:44 pm

i have an index and want to work out an optimized basket from a some futures. How do I account for the variation in the bid/asl spreads of the futures when deciding the optimzed basket. eg suppose there are tow futures I consider to hedge the index. One tracks the index but has a high bid/ask spread. The other future less well but has a tighter spread. Is there some way i can capture this trade off mathematically to decide what is the optimized hedge. NB the futures are use are a proxy for the index. the underlying tracking the futures is not the same as the index.
 
User avatar
endian675
Posts: 0
Joined: March 11th, 2008, 7:09 pm

optimized basket

April 6th, 2010, 1:13 pm

just a quick jot down of my thoughts, take/discard them as you like* tracking error is a function of time, as is the bid-ask spread of the other future. So you want to backtest both to see what's happened in the past* a wide spread can imply low tradeable volume, so if you want to do big size then bear that in mind* how stable is the improvement in tracking error for the wider spread future? does it diverge during market dislocations or periods of high volatility? I suspect it might, and so the future with the "worse" tracking error might prove a better overall hedgeapologies if I'm wildly off base
 
User avatar
winstontj
Posts: 0
Joined: April 7th, 2010, 1:00 pm

optimized basket

April 8th, 2010, 1:57 pm

QuoteOriginally posted by: playeri have an index and want to work out an optimized basket from a some futures. How do I account for the variation in the bid/asl spreads of the futures when deciding the optimzed basket. eg suppose there are tow futures I consider to hedge the index. One tracks the index but has a high bid/ask spread. The other future less well but has a tighter spread. Is there some way i can capture this trade off mathematically to decide what is the optimized hedge. NB the futures are use are a proxy for the index. the underlying tracking the futures is not the same as the index.Are you saying "wide" because the contract you are looking at has a 25-cent tick and the equities in the index have a one-penny spread?Do you have the data and weightings to properly price the index?
 
User avatar
player
Topic Author
Posts: 0
Joined: August 5th, 2002, 10:00 am

optimized basket

April 13th, 2010, 7:28 am

the point is that on average some of the futures can trade with 50 bps point spread. some of the others futures trade with say 10 bps spread. but the 50 points spread ahs lower tracking erro than the 10 bps spread. How can i capture this trade off when doing the regression??
 
User avatar
daveangel
Posts: 5
Joined: October 20th, 2003, 4:05 pm

optimized basket

April 13th, 2010, 7:31 am

QuoteOriginally posted by: playerthe point is that on average some of the futures can trade with 50 bps point spread. some of the others futures trade with say 10 bps spread. but the 50 points spread ahs lower tracking erro than the 10 bps spread. How can i capture this trade off when doing the regression??have you tried optimisation with a constraint ?
knowledge comes, wisdom lingers
 
User avatar
endian675
Posts: 0
Joined: March 11th, 2008, 7:09 pm

optimized basket

April 13th, 2010, 12:53 pm

quadprog is your friend on this one I reckon