[R-SIG-Finance] index tracking

[Patrick Burns]

Forgive my skepticism, but I doubt
that any stocks are really cointegrated
with an index.  The cointegration
approach might work, but I think you'd
be finding highly correlated stocks.

The traditional way (as far as I know)
is to do an optimization that minimizes
tracking error subject to the number-of-
names constraint that you want.

I do agree that daily rebalancing is
almost surely overkill.  Monthly rebalancing
might even be as well.  It is going to
depend on the tracking error you can
tolerate, and how important not breaking
the tracking error bound is to you.



Patrick Burns
patrick at burns-stat.com
+44 (0)20 8525 0696
http://www.burns-stat.com
(home of "The R Inferno" and "A Guide for the Unwilling S User")

Adams, Zeno wrote:
> 
>>> I'm after suggestions for the design of simple index-tracking models.
> I >>have access to good clean data, and I am trying to cook up
> portfolios >>which replicate indices with a minimum number of long
> positions when >>rebalancing is done daily.
> 
> 
> I think a nice way to replicate an index using only a few long positions
> is to use equities that are cointegrated with the index. An example is
> given in Carol Alexander's (2001) book: "Market Models. A Guide to
> Financial Data Analysis" Ch.12. This can be done in R using Paff's urca
> package. You may also want to look at his book "Analysis of Integrated
> and Cointegrated Time Series with R". If done correctly, you don't even
> have to rebalance on a daily basis.
> 
> Zeno
> 
> 
> EBS European Business School gemeinnuetzige GmbH - Sitz der Gesellschaft: Wiesbaden, Amtsgericht Wiesbaden HRB 19951 - Umsatzsteuer-ID DE 113891213 Geschaeftsfuehrer: Prof. Dr. Christopher Jahns,  Praesident; Prof. Dr. Rolf Tilmes, Dekan; Sabine Fuchs, CMO; Aufsichtsrat: Dr. Hellmut K. Albrecht, Vorsitzender
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
> 
>