We propose a generalized ridge-type estimator for the estimation of regression coefficients in the presence of multicollinearity in the design matrix. The first two exact moments and the expressions for the relative bias, relative mean square error and relative efficiency of the proposed estimator are derived.