> beyond some vague allusions that perhaps racial/ethnic profiling should make a comeback.

It suggests nothing of the sort, and in fact explicitly attempts to establish that none of this is intrinsic to any affinity group. The point isn't that the fraudsters in this case were of a particular ethnicity; it's that they shared an ethnicity, which enables the kind of internal social trust that fraud rings require.

Further, it suggests many clear heuristics that are nothing to do with ethnicity. The suggestions are clear and explicit; they just happen to involve occasionally denying services without proof of already-committed fraud, which you apparently consider unacceptable. But here is an experienced person telling you from expertise, with abundant citations and evidence, that nothing else really works, and furthermore that this is common knowledge in a well established industry specifically devoted to the problem.