As a friend of mine pointed out, how is this possible? Shouldn't fraudulent numbers also follow Benford's law?
If you are aware of it, you make sure your fake data follows Benford's law. Otherwise, your fake data will be 'random', because people assume that is how it is supposed to look.
Random data doesn't follow the law. Fraudulent tax returns are essentially random numbers, hence, detectable.
http://falkenblog.blogspot.com/2008/12/benfords-law-catches-...