Evidence against: https://davidcard.berkeley.edu/papers/mariel-impact.pdf
Edit: the paper you linked mention the effect over wages, not over tax revenue and spending.
My experience with social safety nets in EU is that they have been there when I’ve needed them - when I was a poor student, when I was sick — and now that I’m in a different life phase I pay that forward
Your math makes sense at any given instant, but who is on what side changes if you view it over time
Do you have a citation for that broad brush number for what they take?
If there are 101 people, 100 making 101 sandwich per year and 1 doing nothing, and everyone needs 100 sandwiches per year, 99% are contributing more than they take. This works the other way too.
Already your argument is falling apart. Government spending goes disproportionately to children and the elderly who don't work. Workers, even low skill ones, are almost always a net positive.