llikj: Compute log-Likelihood
In PatrickPfeifferDSc/bite: Bayesian Inference On Treatment Effects For Panel Data

Internal. Computes the negativ log-likelihood for all observations with treatment j

1	llikj(sgmaj, rhoj, j, epsy, mu_xst, start_x, nx, start_y)

`sgmaj, rhoj`	vector of 2Tmax* parameter of the variance/covariance matrix
`j`	treatment
`epsy`	residuals on y
`mu_xst`	mean vector of x*
`start_x`	Tmax x trt matrix, containing indicators for panel time - treatment combinations in data vector
`nx`	number of observations
`start_y`	Tmax x trt matrix, containing indicators for panel time - treatment combinations in outcome vector