Girls are all evil (?)

I remember seeing a mathematical proof once that girls were evil. I think it went something like this.

Girls = Time x Money
Time is Money, (eg. Time = Money)
therefore Girls = Money²

Now, Money is the root of all evil (eg. Money = SQRT(All Evil))
therfore Girls = SQRT(All Evil)²
so Girls = All Evil

Ain't math wonderful?

Now, as anyone who reads this mind-bogglingly boring blog knows, I love girls. They smell nice. And only some of them are evil. Which just goes to show - figures lie, and liars figure. The feats a clever pollster can pull with statistics can bring a childlike wonder and awe to the arithmetically aware. One caught my eye the other day from the Wall Street Journal. (I won't bother linking since it requires a subscription).

The editorial cited a recent IRS study breaking out what groups pay what percentage of federal tax revenues. In particular, to make their point that taxes for the very rich really should be reduced, they noted that the highest earning 0.1% of the population paid 5.06% of the federal tax burden in 1979, and was paying 9.52% as of a couple of years ago.

"Gosh," you say. "I know they're really rich and all, but it doesn't seem fair that their burden has almost doubled in the last 25 years!" The editors agreed, stating that this proves that "the overall tax burden grew more progressive from 1979 to 1999," so these hard working individuals "already bear an outsized share of the American tax burden."

I'm fairly certain this statistic is true, as it isn't hard to verify. But wait - are the richest of us paying a larger share of the national taxes because of the tax code?

No, it's because the very rich are earning a far bigger proportion of the national income. In 1979, this same 0.1% of the very (extremely, ungodly, couldn't-spend-it-all-in-one-lifetime-even-if-I-tried) rich took home about three percent (3%) of the national income, and paid about five percent (5%) of the taxes. In 1999, they earned about ten percent (10%) of the national income and paid about eleven percent (11%) of the taxes (as opposed to the again shrinking 9.5% of the taxes today noted in the editorial).

Not only that, but the tax rate for these folks has gone down dramatically. In 1979, the top 0.1% paid, on average, 32% of their income in taxes. Today, they pay less than 23%. So the reason these top 0.1% percent are now paying a higher share of the tax burden than in 1979 is because their share of the national income is rising faster than their tax rates are falling. (A simliar set of statistics were in the same edition of the journal regarding the current Social Security reform discusions - don't get me started...)

Translation from statistics speak -
The Rich Are Getting Richer (and paying a smaller share of their riches to the common kitty).

Now this isn't really news. Unregulated market economies pretty much end up following inverse power laws (as do many natural distributions over time - a brief technical tutorial here.)

And if I'm just sounding jealous and pissed off, well, I am. I share the opinion of approximately 6 billion other individuals on the planet who believe that "if I had that kind of money, I'd do better things with it." I'm not really a fan of income tax (yes, I like consumption taxes better), and don't really think that raising the tax rates on the super wealthy will help or change anything (primarily because they can afford better accountants than the IRS).

Mostly I'm just pissed off at people not thinking rationally, believing what they read uncritically, and behaving like random molecules in a gas that can be modeled stochastically (which is actually the current model used to describe the rest of the wealth distribution - how does it feel to have all the intelligence and influence of a gas molecule? I know, I know, you don't give a fart.)

I could rant for a bit about some filters one should use to evaluate the validity of any statement that uses statistics as it primary support. Survey's are a particular favorite peeve.

• How was the surveyed population selected? What percentage of the population actually responded? What is the size of the population?



• Who sponsored the survey? Who summarized the results? Is the raw data available? How were the questions phrased?



• And if you believe the data - does it matter? Can you do anything about it?

Here's a statement that has been proven true in survey after survey.

In a poll taken among (pick your audience), it was found that 95% of cancer victims had breathed in air, in most cases multiple times each day. In fact, it appeared that the more breaths the individual took during their life, the more likely it was that they would die of cancer.

(People who took fewer overall breaths of air during their lifetime were found to have died of other causes).

This relationship between breathing and cancer is a grave concern, and one worthy of tax investment to find a cure.

(BTW - the poll had a 5% margin of error. See? I'm providing the additional detailed technical information on the poll so you know how really really valid it is!)

I could rant. Yup. But what's the point, really? Someone else could just prove I'm wrong - and have the statistics to back it up.