ks test to compare manager alphas.

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

ks test to compare manager alphas.

Chiquoine, Ben
Hi,

I've recently discovered ks.test() in the stats package.  I'd like to apply it to alphas generated by hedge fund managers. I have several years of monthly excess returns for each manager and I'd like to see which come from distributions that are not statistically different.  I have two questions for the group.  The first question is whether the KS test is valid for two returns series which are likely not iid?  I think I read that as long as the difference in empirical distributions functions (ecdf) is iid from the true distribution (which I don't know) the test is valid.  My second question is more focused on R.  ks.test gives me the source code for the R function but there are several calls to C code within the function.  I'm wondering if there is any way to view the source for the C functions?  I've searched R help for an answer to this less finance related question but only found one post in which they said the source could be found in a file called ks.c.  Unfortunately I!
  can't find this file on my hard drive or on the web.  Any help this group can offer is greatly appreciated.

Thanks,

Ben

___________________________________________
This message and any attached documents contain
information which may be confidential, subject to
privilege or exempt from disclosure under applicable
law. These materials are solely for the use of the
intended recipient. If you are not the intended
recipient of this transmission, you are hereby
notified that any distribution, disclosure, printing,
copying, storage, modification or the taking of any
action in reliance upon this transmission is strictly
prohibited. Delivery of this message to any person
other than the intended recipient shall not
compromise or waive such confidentiality, privilege
or exemption from disclosure as to this
communication.

If you have received this communication in error,
please notify the sender immediately and delete
this message from your system.

        [[alternative HTML version deleted]]

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should go.
Reply | Threaded
Open this post in threaded view
|

Re: ks test to compare manager alphas.

Matthieu Stigler
Le 20. 12. 10 22:04, Chiquoine, Ben a écrit :
> Hi,
>
> I've recently discovered ks.test() in the stats package.  I'd like to apply it to alphas generated by hedge fund managers. I have several years of monthly excess returns for each manager and I'd like to see which come from distributions that are not statistically different.  I have two questions for the group.  The first question is whether the KS test is valid for two returns series which are likely not iid?  I think I read that as long as the difference in empirical distributions functions (ecdf) is iid from the true distribution (which I don't know) the test is valid.  My second question is more focused on R.  ks.test gives me the source code for the R function but there are several calls to C code within the function.  I'm wondering if there is any way to view the source for the C functions?  I've searched R help for an answer to this less finance related question but only found one post in which they said the source could be found in a file called ks.c.  Unfortunately I!
>    can't find this file on my hard drive or on the web.  Any help this group can offer is greatly appreciated.
you are most likely to find the file you search downloading the sources:
http://stat.ethz.ch/CRAN/sources.html
then (at least in R 2.9 sources)
src/library/stats/src

> Thanks,
>
> Ben
>
> ___________________________________________
> This message and any attached documents contain
> information which may be confidential, subject to
> privilege or exempt from disclosure under applicable
> law. These materials are solely for the use of the
> intended recipient. If you are not the intended
> recipient of this transmission, you are hereby
> notified that any distribution, disclosure, printing,
> copying, storage, modification or the taking of any
> action in reliance upon this transmission is strictly
> prohibited. Delivery of this message to any person
> other than the intended recipient shall not
> compromise or waive such confidentiality, privilege
> or exemption from disclosure as to this
> communication.
>
> If you have received this communication in error,
> please notify the sender immediately and delete
> this message from your system.
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should go.
Reply | Threaded
Open this post in threaded view
|

Re: ks test to compare manager alphas.

Patrick Burns-2
In reply to this post by Chiquoine, Ben
I think there are some practical problems.

The ks.test will test if funds are different.
It doesn't really test if one fund is better
than the other.

For the data that you are likely to have, I
doubt you have much power.  I would try
simulating with some known distributions to
see how much power there is.

On 20/12/2010 21:04, Chiquoine, Ben wrote:
> Hi,
>
> I've recently discovered ks.test() in the stats package.  I'd like to apply it to alphas generated by hedge fund managers. I have several years of monthly excess returns for each manager and I'd like to see which come from distributions that are not statistically different.  I have two questions for the group.  The first question is whether the KS test is valid for two returns series which are likely not iid?  I think I read that as long as the difference in empirical distributions functions (ecdf) is iid from the true distribution (which I don't know) the test is valid.  My second question is more focused on R.  ks.test gives me the source code for the R function but there are several calls to C code within the function.  I'm wondering if there is any way to view the source for the C functions?  I've searched R help for an answer to this less finance related question but only found one post in which they said the source could be found in a file called ks.c.  Unfortunately
 I!

>    can't find this file on my hard drive or on the web.  Any help this group can offer is greatly appreciated.
>
> Thanks,
>
> Ben
>
> ___________________________________________
> This message and any attached documents contain
> information which may be confidential, subject to
> privilege or exempt from disclosure under applicable
> law. These materials are solely for the use of the
> intended recipient. If you are not the intended
> recipient of this transmission, you are hereby
> notified that any distribution, disclosure, printing,
> copying, storage, modification or the taking of any
> action in reliance upon this transmission is strictly
> prohibited. Delivery of this message to any person
> other than the intended recipient shall not
> compromise or waive such confidentiality, privilege
> or exemption from disclosure as to this
> communication.
>
> If you have received this communication in error,
> please notify the sender immediately and delete
> this message from your system.
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions should go.
>

--
Patrick Burns
[hidden email]
http://www.burns-stat.com
http://www.portfolioprobe.com/blog
twitter: @portfolioprobe

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should go.
Reply | Threaded
Open this post in threaded view
|

Re: ks test to compare manager alphas.

Eric Zivot
A better method would be to use the bootstrap technique described by Wolfe
and Wunderli. This explicitly controls for the issue of multiple testing.
Similar techniques for controlling data snooping biases are implemented
ttrTest package.
http://www.iew.uzh.ch/wp/iewwp445.pdf


-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Patrick Burns
Sent: Tuesday, December 21, 2010 1:38 AM
To: [hidden email]; [hidden email]
Subject: Re: [R-SIG-Finance] ks test to compare manager alphas.

I think there are some practical problems.

The ks.test will test if funds are different.
It doesn't really test if one fund is better
than the other.

For the data that you are likely to have, I
doubt you have much power.  I would try
simulating with some known distributions to
see how much power there is.

On 20/12/2010 21:04, Chiquoine, Ben wrote:
> Hi,
>
> I've recently discovered ks.test() in the stats package.  I'd like to
apply it to alphas generated by hedge fund managers. I have several years of
monthly excess returns for each manager and I'd like to see which come from
distributions that are not statistically different.  I have two questions
for the group.  The first question is whether the KS test is valid for two
returns series which are likely not iid?  I think I read that as long as the
difference in empirical distributions functions (ecdf) is iid from the true
distribution (which I don't know) the test is valid.  My second question is
more focused on R.  ks.test gives me the source code for the R function but
there are several calls to C code within the function.  I'm wondering if
there is any way to view the source for the C functions?  I've searched R
help for an answer to this less finance related question but only found one
post in which they said the source could be found in a file called ks.c.
Unfortunately
 I!
>    can't find this file on my hard drive or on the web.  Any help this
group can offer is greatly appreciated.

>
> Thanks,
>
> Ben
>
> ___________________________________________
> This message and any attached documents contain
> information which may be confidential, subject to
> privilege or exempt from disclosure under applicable
> law. These materials are solely for the use of the
> intended recipient. If you are not the intended
> recipient of this transmission, you are hereby
> notified that any distribution, disclosure, printing,
> copying, storage, modification or the taking of any
> action in reliance upon this transmission is strictly
> prohibited. Delivery of this message to any person
> other than the intended recipient shall not
> compromise or waive such confidentiality, privilege
> or exemption from disclosure as to this
> communication.
>
> If you have received this communication in error,
> please notify the sender immediately and delete
> this message from your system.
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
should go.
>

--
Patrick Burns
[hidden email]
http://www.burns-stat.com
http://www.portfolioprobe.com/blog
twitter: @portfolioprobe

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions
should go.

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should go.
Reply | Threaded
Open this post in threaded view
|

Re: ks test to compare manager alphas.

Chiquoine, Ben
Eric and Pat,

Thanks for the suggestions. I found the paper by Wolfe and Wunderli to be very clearly written and easy to understand however I'm not sure it's the right answer to my problem. I hadn't intended the ks test to be  a tool which would tell me which managers were the best. What I was actually looking for was a test to see which managers were similar and which were different.  This could provide a useful answer to questions like "Do I really want to invest with a new manager whose returns are extremely similar to a manger that is already in my portfolio?" and "Are a managers claims that they are different from other managers w/ the same strategy enforced by examining the historical return distributions?" My guess is that the ks test's indifference to the correlation between two managers may make it less useful than I'd hoped but anecdotal evidence based on 100 months of historical return data seems to suggest that it may have some value.

Thanks,

Ben
 
-----Original Message-----
From: Eric Zivot [mailto:[hidden email]]
Sent: Tuesday, December 21, 2010 12:46 PM
To: 'Patrick Burns'; [hidden email]; Chiquoine, Ben
Subject: RE: [R-SIG-Finance] ks test to compare manager alphas.

A better method would be to use the bootstrap technique described by Wolfe
and Wunderli. This explicitly controls for the issue of multiple testing.
Similar techniques for controlling data snooping biases are implemented
ttrTest package.
http://www.iew.uzh.ch/wp/iewwp445.pdf


-----Original Message-----
From: [hidden email]
[mailto:[hidden email]] On Behalf Of Patrick Burns
Sent: Tuesday, December 21, 2010 1:38 AM
To: [hidden email]; [hidden email]
Subject: Re: [R-SIG-Finance] ks test to compare manager alphas.

I think there are some practical problems.

The ks.test will test if funds are different.
It doesn't really test if one fund is better
than the other.

For the data that you are likely to have, I
doubt you have much power.  I would try
simulating with some known distributions to
see how much power there is.

On 20/12/2010 21:04, Chiquoine, Ben wrote:
> Hi,
>
> I've recently discovered ks.test() in the stats package.  I'd like to
apply it to alphas generated by hedge fund managers. I have several years of
monthly excess returns for each manager and I'd like to see which come from
distributions that are not statistically different.  I have two questions
for the group.  The first question is whether the KS test is valid for two
returns series which are likely not iid?  I think I read that as long as the
difference in empirical distributions functions (ecdf) is iid from the true
distribution (which I don't know) the test is valid.  My second question is
more focused on R.  ks.test gives me the source code for the R function but
there are several calls to C code within the function.  I'm wondering if
there is any way to view the source for the C functions?  I've searched R
help for an answer to this less finance related question but only found one
post in which they said the source could be found in a file called ks.c.
Unfortunately
 I!
>    can't find this file on my hard drive or on the web.  Any help this
group can offer is greatly appreciated.

>
> Thanks,
>
> Ben
>
> ___________________________________________
> This message and any attached documents contain
> information which may be confidential, subject to
> privilege or exempt from disclosure under applicable
> law. These materials are solely for the use of the
> intended recipient. If you are not the intended
> recipient of this transmission, you are hereby
> notified that any distribution, disclosure, printing,
> copying, storage, modification or the taking of any
> action in reliance upon this transmission is strictly
> prohibited. Delivery of this message to any person
> other than the intended recipient shall not
> compromise or waive such confidentiality, privilege
> or exemption from disclosure as to this
> communication.
>
> If you have received this communication in error,
> please notify the sender immediately and delete
> this message from your system.
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> [hidden email] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
should go.
>

--
Patrick Burns
[hidden email]
http://www.burns-stat.com
http://www.portfolioprobe.com/blog
twitter: @portfolioprobe

_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions
should go.

___________________________________________
This message and any attached documents contain
information which may be confidential, subject to
privilege or exempt from disclosure under applicable
law. These materials are solely for the use of the
intended recipient. If you are not the intended
recipient of this transmission, you are hereby
notified that any distribution, disclosure, printing,
copying, storage, modification or the taking of any
action in reliance upon this transmission is strictly
prohibited. Delivery of this message to any person
other than the intended recipient shall not
compromise or waive such confidentiality, privilege
or exemption from disclosure as to this
communication.

If you have received this communication in error,
please notify the sender immediately and delete
this message from your system.


_______________________________________________
[hidden email] mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. If you want to post, subscribe first.
-- Also note that this is not the r-help list where general R questions should go.