In the paper, a study of target localization performances is presented for coherent multiple-input multiple-output (MIMO) and single-input multiple-output (SIMO) radars systems with widely separated elements. The evaluation is based on the best linear unbiased estimator (BLUE), providing the localization mean squared error (MSE) in a closed-form solution. This estimator elucidates the relation between the radar locations, target location, and localization accuracy through the use of the geometric dilution of precision (GDOP) metric. Contour maps of the GDOP relate a given deployment of sensors and the achievable accuracy to the at various target locations. This metric is shown to represent the spatial advantage of the system. The best achievable accuracy for both configurations is derived. MIMO radar systems with coherent processing are shown to benefit from higher spatial advantage, compared with SIMO systems. The advantage of the MIMO radar scheme over SIMO is evident when considering the achievable accuracy for a radar system with M transmitters and N receivers, rather than 1 transmitter and MN receivers. It is shown that MIMO radar, with a total of M +N sensors, has twice the performance (in terms of localization MSE) of a system with (MN + 1) sensors.