Error when increasing size of dependent variable linear regression

Suppose I have the following forloop which estimates parameters of a linear model

count = 0;

nmonth = 5;

for j = 1:dimensions(2) % number of stocks 1 , 2 and 3

for i = 1:nmonth

mdl = fitlm(X((30*i-30+1):i*30,:),Y((30*i-30+1):i*30,j)); % regression

% for one month of the regressors X on Y for stock 1.

count = count+1;

montherrors(:,(count)) = mdl.Residuals(:,1); % daily errors for

errors = table2array(montherrors); % convert to array so I can perform

% calculations

stdev = std(errors,1);

idioallstocks = stdev * sqrt(ndays);

end

end

Warning: Regression design matrix is rank deficient to within machine precision.

Now when dimensions(2) is for example 50 it works perfectly fine. But when dimensions(2) is for example 100 I get the error Warning: Regression design matrix is rank deficient to within machine precision and I do not know why.

[Torsten]

Maybe there are NaN / Inf values in the Y-matrix for j > 50 ?

[Wietze Zijpp]

Thank you for this response !