Cheat...... if you Know How

In Everybody Lies - an interesting look at big data in the Internet age - the author shares a data study from one of his mentors who looked at reams of IRS data.

The study examined at how much single, self-employed people with one child claimed on their taxes and found that the most popular earnings number was $9,000. It just so happens that $9,000 is the point where the claimant gets the biggest refund. When examining audits, almost all these returns earned less or more than $9,000; some wildly so.

Although that's interesting in itself perhaps, the study looked at the data a little more deeply.

The researchers subset these $9k claims and found there was some serious variance between locales. In Philly, for example, 2% triggered a positive, Miami, 30%. When they ran this data for demographics, income and other factors, two things stuck out. First, the higher the incidence of self employed people in an area, the much higher the likelihood $9k (exactly) would be claimed. Second, the more tax professionals in the area, the higher the incidence of the $9k number.

They conclude if you had i) knowledge and ii) friends, or professionals nearby to spread that knowledge to you, you were likely to know the magic number; the best number to cheat at. The authors contend it wasn't really about wanting to cheat, or being bad people. It was an effect of what marketers call "contagion".

This helps, I think, explain cheating in horse racing. If there are a lot of guys and gals pushing the envelope at your favorite track (we won't name names), it might not be by chance. It might be because knowledge of how to cheat is more readily available. It might be because (some) vets - like tax professionals - are there to spread knowledge of what they can do for you.

For regulators, this probably helps explain why it's smart for them to concentrate on certain tracks, barns and vets. If they stop one, the contagion spreads in reverse, and they end up stopping many.

Have a nice Wednesday everyone.