Problem with missing data in a time series

Hello, I need a hand on this problem: In an Excel workbook I reported 10 time series (with monthly frequency) of 10 titles that should cover the past 15 years. Unfortunately, not all titles can cover the 15-year time series. For example, a title only goes up to 2003; So in the column of that title, I have the first 5 years with a "Not Available" instead of a value. Once I’have imported the data into Matlab, obviously, in the column of the title with the shorter series appears NaN where there are no values.

Prices xlsread = ('PrezziTitoli.xls')

whos

% Name Size Bytes Class Attributes
%
Prices 182x10% 6360 double

My goal is to estimate the variance-covariance matrix, however, because of the lack of data, the calculation is not possible for me. I thought to an interpolation, before the calculation of the variance-covariance matrix, to cover the values that in Matlab return NaN, for example with a "fillts", but have difficulties in its use.

Can you help?

Thanks

Answer 1

[Shashank Prasanna]

You can use NONCOV to compute the covariance matrix for data with NaN:

http://www.mathworks.com/help/stats/nancov.html

[Fabrizio Marinelli]

Hi, this could be a good idea! The problem is that i have to calculate also the log-returns, and for that i need full columns in the prices series.

[Shashank Prasanna]

Then interpolation is way to go. What is the problem with using fillts? It does require you to create a financial time series object first. If you prefer to just use interp1, check the doc of interp1:

http://www.mathworks.com/help/matlab/ref/interp1.html Unless you provide some code you tried, it will be hard to help.

[Fabrizio Marinelli]

Thank you, thank you! I solved.. Following your first advice I calculated the log-returns in normal way and than i used nonmean, nonvar, etc. for the statistics, excluding NaN's rows.

Answer 2

[Titus Edelhofer]

Hi,

this answer is more of a general advice rather than answering the question, sorry for that ;-).

I'm not sure that it will make sense to try to go from 2003 back 5 years "guessing" the data. I would assume it makes more sense to restrict yourself to the last 10 years, i.e., where you have all the data and compute the variance-covariance matrix for this time period only ...

Titus

[Fabrizio Marinelli]

Hi, I know that it would be better to reduce the period of observation. Unfortunately, because of a project for the university, I must act over 15 years. Anyway thanks for the suggestion.