The role and construction of Stein-rule estimators in multivariate ultrastructural
model is discussed when some prior information about the regression coecients
is available in the form of exact linear restrictions. The additional information
in the forms of covariance matrix of measurement errors and reliability matrix
of explanatory variables is used for the construction of consistent estimators.
Two families of Stein-rule estimators are proposed using each type of additional
information which are consistent as well as satisfy the exact linear restrictions.
The distribution of measurement errors is assumed to be not necessarily normally
distributed. The asymptotic distribution of the proposed families of Stein-rule
estimators are derived and studied. The nite sample properties of the estimators
are studied through a Monte-Carlo simulation experiment.