This paper addresses the problem of vehicle position estimation in dense urban environments, where traditional global positioning system (GPS)-based localisation techniques are severely affected by non line-of-sight (NLOS) signal propagation and multipaths presence. Assuming that GPS signals are fairly received only by a very small fraction of vehicles at the border of the urban environment, we propose the multihop scheme based on vehicle to X (V2X) communication to propagate this information to the whole network. As a consequence, the multihop scheme allows every vehicle to estimate its own position collecting its cumulative distances to border vehicles. Finally we introduce a new analytical framework to verify the fundamental performance of multihop scheme for location forwarding in urban dense environments in terms of position estimate error bounds, jointly considering the uncertainty introduced by the multihop process and by the GPS localization.