It's perhaps worth pointing out that they have 2020 at 1% right now, but they also mentioned how they raised 2019 from 1.1% to 2.1% due to new discoveries (your

It's perhaps worth pointing out that they have 2020 at 1% right now, but they also mentioned how they raised 2019 from 1.1% to 2.1% due to new discoveries (your first source mentions this). I question how effectively we can really quantify this if the number can double a year after being reported, at the very least it's hard to say how accurate the initial numbers are. It's probably more accurate to say 1% is a lower bound for 2020, but a lower bound is obviously not that useful. Perhaps it would be better to try and gather the amount of crypto use that is definitely legitimate, since that would at least give a theoretical upper-bound. Give the assumption that it's such a larger group that may just be too much to work through.