Abstract The sensitivity of the midlatitude storm track and eddy-driven wind to the sea surface temperature (SST) boundary forcing is studied over a wide range of perturbations using both simple and comprehensive general circulation models over aquaplanet lower boundary conditions. Under the single-jet circulation regime similar to the conditions of the present climate in the Northern Hemisphere winter or the Southern Hemisphere summer, the eddy-driven jet shifts monotonically poleward with both the global mean and the equator-to-pole gradient of the SST. The eddy-driven jet can have a reverse relationship to the gradient if it is well separated from the subtropical jet and Hadley cell boundary in a double-jet circulation regime. A simple scaling is put forward to interpret the simulated sensitivity of the storm-track/eddy-driven westerly wind position within the single-jet regime in both models. The rationale for the scaling is based on the notion that the wave activity flux can propagate horizontally away from the source region, resulting in a broader distribution of eddy potential vorticity (PV) flux in the upper troposphere than that of the flux in the opposite direction in the lower troposphere. As a consequence, the position of the maximum of the eddy-driven westerlies tends to be controlled by the profile of the relatively sharp-peaked low-level PV flux, which is dominated by the eddy heat flux component of the Eliassen–Palm (EP) flux. Thus, the position of the eddy-driven surface westerlies may be inferred from the vertical EP flux coming out of the lower troposphere. The vertical EP flux can be parameterized by a measure of baroclinicity, whose latitudinal variations show a linear relationship with the meridional displacement of the eddy-driven westerlies and the storm track. This relationship still holds well within the single-jet regime, even when only the variation of static stability is taken into consideration in estimating the baroclinicity (the temperature gradient component of which is fixed). To the extent that the static stability is deterministically constrained by and hence can be predicted from the given SST conditions through a moist scaling for the midlatitude stratification, one may, given SST perturbations, predict which way the storm track and eddy-driven wind should shift with respect to a chosen reference climate state. The resultant anomaly-wise scaling turns out to be valid for both the idealized and comprehensive models, regardless of the details in the model physics. By corollary, it can be argued that the poleward shift of storm track found in the global warming simulations by fully coupled climate models may be attributed, at least partially, to the increase in the subtropical and midlatitude static stability with global warming.