I guess it's just a drive to reduce uninsured modifications. I suspect they analysed the data on policies taken out without mods where quotes were made with mods, versus policies with no mods where they found mods to be fitted for whatever reason, and found a correlation.