In this paper we provide a method to analytically compute the energy saving provided by the use of Transmission Power Control (TPC) at the MAC layer in Wireless Sensor Networks (WSN). We consider a classical TPC mechanism: data packets are transmitted with the minimum required power to achieve a given packet error probability, whereas the additional MAC control packets are transmitted with the nominal (maximum) power. This scheme has been chosen because it does not modify the network topology, since control packet transmission range does not change. This property also allow us to compute analytically the expected energy savings. Besides, this type of TPC can be implemented in the current sensor hardware, and can be applied directly to several MAC protocols already proposed for WSN. The foundation of our analysis is the evaluation of L ratio, defined as the total energy consumed by the network using the original MAC protocol divided by the total energy consumed if the TPC mechanism is employed. In the L computation we emphasize the basic properties of sensor networks. Namely, the savings are calculated for a network that is active a very long time, and where the number of sensors is supposed to be very large. The nodes position is assumed to be random -for the sake of example a normal bivariate distribution is assumed in the paper- and no node mobility is considered. In the analysis we stress the radio propagation and the distribution of the nodes in the network, that will ultimately determine the performance of the TPC. Under these conditions we compute the mean value of L. Finally, we have applied the method to evaluate the benefits of TPC for TDMA and CSMA with two representative protocols, L-MAC and S-MAC using their implementation reference parameters. The conclusion is that, while S-MAC does not achieve a significant improvement, LMAC may reach energy savings up to 10-20%.