That may be because of magnetic fields from the current flow through the plasma disturbing the Hall Sensor in your clamp-meter. See it it disturbs a compass-needle. Or if you don't have one, magnetise an ordinary sewing needle and suspend it by the balance point from a piece of thread.
Here is a 10x sped-up gif animation of an old compass used in a similar test. There seems to be a response, but the movement is so small that it's difficult to tell for sure which is the real cause. I didn't notice it until I accelerated the video.