I bet that if you actually work out what's happening, at best it's equivalent to just emitting photons, which you can do by just aiming a flashlight out the back.
"Efficiency" isn't a well-defined quantity, though. What you are looking for is thrust/power, and even a device that converts energy to photons with 100% efficiency produces preciously little thrust per power used.
Conservation of momentum dictates that the momentum gain follows from the net momentum of emitted photons, so it really should be simply a matter of shinning light away. dp=dE/c (and since magnets have histeresis losses as mentioned, they radiate energy isotropically due to heat, lowering efficiency).
Ya that's definitely true. You might check out JLN labs, they have a bunch of over unity and reactionless rocket experiments drawing from the work by TT Brown:
Some of the more fascinating ones are in near-vacuum. I remember reading that asymmetric fields might be able to propel virtual pairs or interact with the interstellar medium, but unfortunately the effect seems to fall too quickly in relation to density.
There was another article about how since cooper pairs are delocalized in a superconductor, they may not interact with gravity. It was trying to explain things like the electrogravitic effects near spinning superconductors:
I keep thinking that if any of these do stray from physics, it will be interesting to see how they reconcile it. So far it hasn't happened but days like today make me wonder if it's just around the corner.
"Efficiency" isn't a well-defined quantity, though. What you are looking for is thrust/power, and even a device that converts energy to photons with 100% efficiency produces preciously little thrust per power used.