In many cases the perfect capital mobility hypothesis is used by default. Therefore, if we follow this idea there should not be any connection between internal savings and investments in a country or at least this connection must be not significant. But some of empirical research demonstrate opposite results. In economic literature this concept is well-known as the “Feldstein — Horioka Puzzle”. Considering the relative lack of studies concerning this theme on the regional level, it seems quite perspective to research the hypothesis from the mesoeconomical point of view. So, the central aim of this article is the verification of regional capital mobility in the Russian Federation, following the logics of the Feldstein — Horioka Puzzle. The main peculiarity of the paper is based on the application a spatial econometric approach. The authors use SLX (spatial lag of X model) and SDM (spatial Durbin model) models to achieve the aim. As the result it was found out that there is almost perfect capital mobility situation on the regional level in the Russian Federation. Hereby the Feldstein — Horioka Puzzle was rejected. Along with it several significant direct and indirect effects were established (for instance, reinforced interaction between geographical neighbors-region, absent of capital flow from some regions and other). The results can be used for the consideration of relatively more efficient regional government policy.