Let y1, . . , yn be random variables with var(yi) = σ 2 and cov(yi , yj ) = rhoσ2 . the variance covariance matrix is of the form (a − b)i + bj, where a = 1, b = rho, j = 110 . therefore, in our model σ = σ 2 [(1 − rho)i + rhoj]. the inverse of this special matrix can be obtained as follows: σ−1 =