Roy Lewallen wrote:

Broadband transformers, which can operate well over several decades of
frequency, commonly use ferrite cores which are essentially resistive
over most of the operating frequency range. The sign of the impedance is
unimportant to the transformer's operation; all that's necessary is that
its magnitude be adequately high over the operating range (and of course
that the core's permeability be adequately high). The wide band high
impedance requirement is virtually impossible to meet with an inductive
core whose impedance is approximately proportional to frequency, but
easily done with cores whose impedance is essentially resistive.

As long as the impedance is high enough, what is the need for it to be
constant (or even nearly so) over a useful frequency range?