A generalized ridge-type estimator for the estimation of regression coefficients is proposed in the presence of multicollinearity in the design matrix. The first two exact moments and the expressions for the Relative Bias. (RB), Relative Mean Square Error (RMSE) and Relative Efficiency (RE) of this estimator are derived. A numerical comparison of relative efficiency of this estimator with respect Ordinary Least Square Estimator (OLSE) and the studied ridge estimator is also presented.