“…The idea is simple [Kakushadze, 2015]. It is based on the observation that, as K approaches M , min(ξ 2 i ) goes to 0 (i.e., less and less of the total variance Γ ii ≡ 1 is attributed to the specific variance, and more and more of it is attributed to the factors), while as K approaches 0, max(ξ 2 i ) goes to 1 (i.e., less and less of the total variance is attributed to the factors, and more and more of it is attributed to the specific variance).…”