Attempt to grow array along ambiguous dimension - loop error

I am trying to calculate standard errors in a loop of regressions that I am performing on mutual funds however when i try to run my code o get the 'Attempt to grow array along ambiguous dimension' error message.

Here is my coding so far:

 [r,c]=size(FR);
 alpha=zeros(c,1);
 x=FR(:,i);
 y=isnan(x); 
 sy=sum(y); 
 x=x((sy+1):end); 
 lx=length(x); 
 X=zeros(lx,5);
 X(:,1)=ones(lx,1);
 X(:,2:5)=C4F((sy+1):end,:);
 beta=inv(X'*X)*X'*x; 
 alpha(i)=beta(1);
 X(:,1) = alpha(lx,1);
 CM(i) = cov(X(i));
 SE(i) = diag(sqrt(CM(i)));

 end

I know that the error first relates to the line 'CM(i) = cov(X(i));' however I am unsure what other way there is to obtain the covariance matrix for every fund in my sample (1600 funds). Does anyone know a way to achieve this or correct my current problem?

Thanks in advance!

[Geoff Hayes]

Daniel - I think that you have omitted some of your code. Do you have a for or while loop? What is CM initialized to? What are the dimensions of cov(X(i)) as it might not be a scalar that you are assuming it to be...