Dealing with missing data and consistent parameters

Discussion on Quantitative Methods in Finance
Forum rules
DISCLAIMER: We do not warrant or represent that this forum or its content is free of viruses, worms or other code that might be contaminating or destructive. We cannot guarantee that documents or files downloaded from the Site will be free from viruses and we do not accept any responsibility for any damage or loss caused by any virus. Accordingly, for your own protection, you must use virus-checking software when using the forum. You must not post or provide to us via the forum, any document or file which you believe may contain a virus. You must virus check any document or file which you intend to post or provide to us via the forum. You must ensure that any document or file you intend to post to the forum does not contravene any applicable laws or contravene any person's legal rights. We do not accept any responsibility for any damage or loss you may suffer.
Searle
Posts: 2
Joined: Fri Jun 26, 2009 7:24 am

Dealing with missing data and consistent parameters

Postby Searle » Sun Jun 23, 2013 3:58 am

Hi Prof Alexander,

I hope this finds you well.

I would like to run a few questions by you and hear your thoughts if you have some time.

Firstly, when doing Monte Carlo simulations for correlated processes, is it essential that the time frame used to calculate the individual volatilities corresponds to the time frame used to calculate the correlations? On a related note, is there any inconsistency if, when doing this, the volatilities used in the simulation are, say, EWMA, which differ to those used to calculate the standard (linear) correlations? (The following link is also of benefit http://www.carolalexander.org/publish/d ... r_2008.pdf.)

Secondly, when dealing with two or more time series with missing data (here, I am not referring to holidays etc., but rather a lack of liquidity), is it best to limit the data to only those dates where there is common data? Alternative approaches entail filling in for missing data by using the previous value(s) (which I think will bias the estimations), or to use some sort of data filling technique (E.g. EM). Another approach could entail using Kendall’s tau estimation on all the available data.

I would love to hear your thoughts on these questions.

Thanks so much for your time, I really appreciate it as I am sure you are extremely busy.

coalexander
Posts: 815
Joined: Sun Sep 28, 2008 10:30 pm

Re: Dealing with missing data and consistent parameters

Postby coalexander » Thu Jul 25, 2013 8:36 pm

Hi Searle,

Many apologies to you and other forum users for my absence. I have been overwhelmed with work.

Regarding your first question you must always compute vols and corrs using the same data and methods, otherwise your covariance matrix may not be positive definite (giving you negative portolfio volatilieis, amongst other nasty things) and also it would not really mean anything.

I would usually try to just exclude the dates with no data rather than fill in data using interpolation, EM, setting zero return, or some other means. As long as there are only a few dates, this should not be a problem.

Cheers, Carol


Return to “Volume I”

Who is online

Users browsing this forum: No registered users and 1 guest