As you know, the way that the dummy load resistors are wired for an EMC test can actually drastically effect the scan....the common mode noise can radiate off the dummy load wires , and indeed, from the dummy load resistors.
Why did they not choose a more tight layout of the load in DER472, page 65?, ....ie, so as to reduce load current loop areas..because then you would have got a better (lower) EMC scan.
I'm of the opinion that your test article should consist
of only what you deliver, any "add-ons" should be of
a form unlikely to compromise the result -and- their
impact -to- those results should be quantified and a
method to "de-embed" it validated & bought off.
But, you know, go ahead and hang some antennae off
it, if you like to create exercise for yourself.
I wouldn't sign up for the EMC of somebody else's
harness. That's Somebody Else's Problem.