 # Trailing Statistic Data

I am wondering if anyone knows where to get a free program that will provide me with trailing standard deviation and covariance values??

What exactly are you looking for? You can calculate standard deviation and covariance in excel.

You can get a time series of prices easily using Yahoo Finance or similar. From there, calculate the sd and covariance with Excel formulas.

Is this the only way you guys know how to do it? If this is the case and I am looking at a 10 asset portfolio, I am looking at doing quite a bit of covariance calculations. I recall making a covariance matrix during university, but I believe that I used the same time consuming manual process to build the matrix. I would really like some sort of program where I could simply enter the tickers and receive trailing Cov values.

If you have Bloomberg, there is an application that lets you do that: CORREL .

try Bloomberg…i have never used it but i have seen something about std dev. before…

I wrote some code in R that will query Yahoo Finance for price/levels information and dump it into a dataframe.

From there you can transform it to returns (or you can download returns direct from yahoo, but I prefer to start with levels data) and run a correlation/covariance matrix on the most recent N periods.

R has a challenging learning curve, but once you get there, it’s extremely useful and surprisingly powerful for a something that’s basically free.

If you don’t want to do that, you can download data files from Yahoo, copy into excel, and do the correlations there, but it can be a bit clunky to code the covariance matrix. Use of OFFSET is likely to help.

I use an excel VBA macro to download the ticker data from yahoo. Then I use the Excel>Tools>Data Analysis>Correlation function to calculate a correlation. If you wanted to see the trailing standard deviation that would be easy to do in excel, but I don’t know how you’d show the trailing covariance. The covariance is a matrix, so showing it on a trailing basis would require a different matrix for each trailing time period. That seems like it’s going to be messy and hard to interpret. For only 10 stocks, excel seems like it would work fairly well.