Returns and standard deviation groupwise

I have a time series of daily stock prices for several firms. Now I want to calculate the daily returns and the rolling standard deviation of 250 days for each firm. I have no clue how to implement that in Matlab. Whereas in Stata I could only use the "by" command, it didnt find a similar function in Matlab.

I did group my firm ID and tried to apply splitapply. It works but what I get is a cell array. If I do not specify "{}" (cell array) in my function, this function wont work.

func =@(p){log(p(2:end)./p(1:end-1))} ;
r = splitapply(func,p,Entity);

The goal is to calculate daily returns (in an array) and annualized std. dev to further calculate the distance to default for each company.

Has someone a hint, how this might work? Thanks for any hint!

Answer 1

[John D'Errico]

I'm not sure what you are asking by "daily returns". Perhaps you want a moving mean, on a 250 day time window. That is simple.

mean250 = conv(dailydata,ones(250)/250,'valid');

But a moving window standard deviation is also easy enough. Download my movingstd tool from the File Exchange. It uses convolution with a trick to get the std in a window, and do it efficiently.

[Leo]

thanks a lot! My main issue doing this calculation for each firm, such that if I have a time series of firms and their stock prices that the return in t1 is zero and t+1: r+1 = log(pt+1/pt)