I am not expert in this matter and my words are mostly based on bad memory and some logic.
I imagine this stator as a stack of lams already shorted on inside, where they are pressed on the aluminium cage. In this picture everything still looks right, no loops and current paths this far? (only cage loop but it is far from magnetic hell on the toots)
Now you add sand granule ripped out conductive deformed iron pieces, stretched in sanding process bridging lamination on opposite, tooth side, creating large number of possible current paths of different length (R) all across the stator and cage.
It just seems to cancel out all benefits of 0.35mm lamination and lamination at all.
I hope to come across a study, explaining the impact of given situation to cancel out my paranoia.