
This observation is not necessarily true at the liquid boundary. Since no molecules "like" to point at a hydrophobic interface direction (this would imply many uncompensated hydrogen bonds), most of the molecules orient along the liquid surface, the boundary between the liquid and a hydrophobic may become polarized (become essentially ferroelectric). Practically this amounts to a formation of stable network of hydrogen bonds on the interface.

The second graph corresponds to the polarization density (mean dipole moment of a liquid volume). Comparing the two graphs we see that the interface is indeed polarized and the polarization decays quickly into the bulk both of the gas and the liquid phase and, as the detailed calculation shows, contributes to the surface tension coefficient considerably.
In short we observe that depending on the ordering state of the water layer next to a molecular surface the effective surface tension may be different by a large number. Another observation suggests that the water density depletion next to a fully hydrophobic (and thus ordering) surface can be large (up to 30-50%)