There could be occasion for you to make your own IBIS
model for a passive element. Or you could be too literal
minded. If this is (say) about a LVDS or similar physical
layer, and its termination, are you convinced that an IBIS
analysis will show you differential signaling care-abouts
even if you made an I, V correct representation of the
through-resistance?
Or, are you in fact using SPICE with IBIS-derived subcircuit
models (at the end of the tool-chain) in which case a SPICE
resistor might be the done thing?