Asymptotic Efficiency of CCA Subspace Methods in the Case of no Exogenous Inputs

In this paper one of the main open questions in the area of subspace methods is answered partly. One particular algorithm, sometimes termed CCA, is shown to be asymptotically equivalent to estimates obtained by minimizing the pseudo maximum likelihood. Here asymptotically equivalent means, that the difference of the two estimators times the square root of the sample size tends to zero