Handling missing values in predictor when outcome is a Multivariate Normal distribution

Questions v5
2

You would need to impute (give some prior to) those missing x