Economic Populism At The Times

Readers of the Dead Tree Times get a special front-page "looks like she didn't get the memo" treat today.

From a front-pager on some economic news:

Census Reports Slight Increase in ’05 Incomes 

The nation’s median household income rose slightly faster than inflation last year for the first time in six years, the Census Bureau reported yesterday.

The rise, however, had little to do with bigger paychecks — in fact, both men and women earned less in 2005 than 2004. Rather, census officials said, more family members were taking jobs to make ends meet, and some people made more money from investments and other sources beyond wages.

...

That figure fell 5.9 percent between the 2000 census and 2005, to $46,242 from $49,133, according to an analysis of the data conducted for The New York Times by the sociology department of Queens College. The difference was so sharp, in part, because the 2000 census measured 1999 income, which was at the height of the dot-com bubble.

   

Emphasis added.  Elsewhere on the front page we find memo-misser Linda Greenhouse, filling a slow news day with ruminations about hiring practices at the Supreme Court:

...While their pay is a modest $63,335 for their year of service, a Supreme Court clerkship is money in the bank: the clerks are considered such a catch that law firms are currently paying each one they hire a signing bonus of $200,000.

Modest?  I bet that in comparison to other lawyers (and Times readers!), $63K is modest.

MORE:  The Greenhouse article cites this Volokh Conspiracy post from July.  Ms. Greenhouse also puts a puzzle before the statisticians in the house:

In interviews, two of the justices, David H. Souter and Stephen G. Breyer, suggested that the sharp drop in women among the clerkship ranks reflected a random variation in the applicant pool.

From the story, I glean that there are 37 clerkships, and "normally" about 1/3 go to women.

My coffee-table attempt to apply the normal approximation to the binomial distribution left me with a standard deviation equal to the square root of [(2/9) / 36], which became [(1.4/3)/6], which became .5/6, or .08.

From that, I conclude that 95% of outcomes should show 33% +/- 16%, or 1/3 +/- 1/6; call it a range of 16% to 50%.

Seven clerks out of thirty-seven is at the bottom of (but within) the random range.  But I only had one cup of coffee and no calculator when I did that (No, I have no current excuse for the lack of a calculator; obviously, I am at my computer now...)

Greg Mankiw had thoughts on the earlier Times article about lagging wages.

Comments

[clarice]

As more people have enough income and desire to invest, I expect more will have investment income. Is this a bad thing? Is only money earned by the sweat of one's brow significant? Hmm, poor Pinch.

[cathyf]

I've always been curious about something in the statistic -- doesn't most of this data come from tax filings?

So, imagine that the Smith family has one breadwinner, who makes $4000 per month. Little Johnny Smith has crooked teeth, and the Smiths are paying the orthdontist $180/month to fix his teeth. They are using a flexible spending account for this.

Before orthodontia:
gross pay: $4000
taxes: $1000
check: $3000

During orthodontia:
gross pay: $3820
taxes: $955
checks: $2865 + $180 = $3045

So, according to the way these statisticians calculate things, did the Smith family income go down from $4000 to $3820 (because the IRS doesn't "see" the flex account money as "income") or does it go up by the $45 -- the difference between the $3045 and $3000 net pay?

This is a classic mechanism in social-program demogoguery. The demogogues produce all these statistics about poor people's incomes, and create political will to create programs to help them -- food stamps, rent subsidies, day-dare subsidies, free school lunches and breakfasts, etc., etc. Then they come back later and produce more statistics on poor people's incomes but they don't count the value of any of these non-cash programs.

It's really stupid and counterproductive that they pull this crap -- you create political will by making people feel guilty, but then you keep flogging the guilt card. Until eventually people get the message -- they are doing all these things and spending all of this money and it's not doing any good. So let's get rid of the programs and spend the money on monuments to Bob "KKKleagle" Byrd instead...

[Chuck Simmins]

I'm on my third year of blogging my analysis of this report. This year's highlights include a record high ratio for parity between wages for men and women. The 2005 overall poverty rate of 12.6% is higher than the rates for all the years from 1980 through 1998. The poverty rate for blacks in 2005 of 24.7% is lower than the rates in the period from 1959 through 1998. The Hispanic poverty rate for 2005 of 21.8% is lower than the rates from 1980 through 1999.

Average Poverty Rate - First Five Years of an Administration:
Reagan: 14.5%
Clinton: 14.1%
Bush 43: 12.3%

http://northshorejournal.org/index.php/2006/08/poverty-in-america-2005-overview

[Foo Bar]

the normal approximation to the normal distribution

I'm sure what you meant to write here was "the normal approximation to the binomial distribution".

Leaving that aside, I am happy to report that I arrive at the same answer. Well done, sir!

[cathyf]

Isn't the binomial distribution the normal approximation of the normal distribution?

(Oooo... Statistics jokes. Two statisticians go hunting. They see a gorgeous huge buck. The first statistician fires 5 feet in front of the deer while the second fires 5 feet behind it. The turn, exuberantly give each other high fives, and say, "Got him!!!")

[Tom Maguire]

I'm sure what you meant to write here was "the normal approximation to the binomial distribution".

Yikes! Well, I am comfortable with an analysis based on "original intent" in this context. Thanks.