The tail currents predicted by empirical magnetic field models and global MHD simulations are compared. It is shown that the near-Earth currents obtained from the MHD simulations are much weaker than the currents predicted by the Tsyganenko models, primarily because the ring current is not properly represented in the simulations. On the other hand, in the mid-tail and distant tail the loba field strength predicted by the simulations is comparable to what is observed at about 50RE distance, significantly larger than the very low lobe field values predicted by the Tsyganenko models at that distance. Ways to improve these complementary approaches to model the actual magnetospheric configuration are discussed. ¿ 1995 American Geophysical Union