I don't think that the problem is so much whether one can physically do a simulated faillure in a lightie.
Rather, as with all exercises, the real concern is whether the IP can rescue the situation from whatever foul up eventuates - whether that be due to the aircraft's lack of performance or the student/pilot under test's error.
ICO simulation at low level just trims your risk fat down to nothing.
Why not do the exercise at 1000 or 1500ft ? When it all goes pear shaped (as it will from time to time) you get a bit more time to catch it and fix the problem before you hit the hard bits. No guarantees .. but the odds improve greatly.
I got around the problem for all tests etc simply by making it VERY clear to the testing officer that a simulated failure below (state a height for the situation that I chose to nominate) WOULD result in both throttles being closed and the aircraft landed ahead.
Strange .. was never put to the test .. and they knew that I wasn't about to muck about.
Stupid stuff ends up, sooner or later, with embarrassment or worse.
Two very experienced test officers I knew well and with whom I periodically flew (including the one cited by A37575 who died of burns after the Camden prang) both were very skilled but ran out of steam when they simultaneously ran out of height, speed, and ideas.
The ADF materially changed a lot of ways it did business following a few very visible prangs, including the dreadful B707 crash off the Victorian coast.
Hero stuff is for simulators .. generally the worst that can happen is that the other guys get to have a good belly laugh at the pilot's errors.
The only folk who have to tread into the boundaries are the certification TPs and they tend to do things very conservatively these days.
Two areas are stupid if travelled without great care .. low level OEI operations and Vmca simulations.
On the latter matter, I recall the way a good colleague would handle the excesses of GA instructors when forced to do an endorsement, say, to complete an IR renewal ... he just applied a counter load on the rudder so that the obtained static Vmc was some considerable margin above the value for full rudder .. had a number of instructors scratching their heads .. "it went a lot slower than this last time ... ?
Having known A37575 for many years I can attest that he is both a very competent and conservative instructor. One would do well to heed his greybeard observations.