Thursday, September 29, 2011

Grading PolitiFact: President Obama and the Buffett fallacy (Updated)

Context matters -- We examine the claim in the full context, the comments made before and after it, the question that prompted it, and the point the person was trying to make.
--Principles of PolitiFact and the Truth-O-Meter

The issue:

(clipped from

The fact checkers:

Louis Jacobson:  writer, researcher
Bill Adair:  editor


President Obama's words, in context (yellow highlights added to indicate the portion quoted by PolitiFact, "laughter" and "applause" notes deleted):
THE PRESIDENT: Now, this isn’t to punish success. What’s great about this country is our belief that anybody can make it. If you’re willing to put in the sweat, if you’re willing to roll up your sleeves, if you’re willing to work hard, you’ve got a good idea, you’re out there taking a risk -- God bless you. You can make millions, you can make billions of dollars in America. This is the land of opportunity. That’s great. All I’m saying is, if you’ve done well -- I’ve done well -- then you should do a little something to give something back. You should want to see the country that provided you with this opportunity to be successful, and be able to provide opportunity for the young people who are going to be coming up behind you.

And all I’m saying is that everything should be fair. You know, you learn the idea of fairness when you’re two, three years old. Right? You’re in the sandbox and you don’t want to let somebody play with your truck and your mom or your daddy go up and they say, “No, hon, that’s not fair, you’ve got to share.” Isn’t that what they say? Things have to be fair. So all I’m saying is that Warren Buffett’s secretary should not be paying a lower tax rate on her income than Warren Buffett.  That doesn’t make any sense. A construction worker who’s making 50 or 60 grand a year shouldn’t be paying higher tax rates than the guy who’s making $50 million a year.  And that’s how it’s working right now. Because they get all these loopholes and tax breaks that you don’t get.

So for me to say, let’s close those loopholes, let’s eliminate those tax breaks, and let’s make sure that everybody is paying their fair share -- there’s nothing wrong with that.
 After noting that Obama's statements contain quite a bit of unverifiable opinion, the fact checkers weigh in on the fact check:

In his Cincinnati speech, Obama did make a comment that was Truth-O-Meter-ready. By saying "and that’s how it’s working now," Obama is saying this pattern of taxation is a common practice.
Though the president's statement qualifies as slightly ambiguous, the PolitiFact judgment seems reasonable on its face:  Obama is saying that a $50,000 earner is routinely taxed at a higher rate than the $50 million earner.  Therefore the president requires more than just a few appropriate individual cases to justify his claim.

We found an IRS chart for tax year 2008 that shows a variety of tax information broken down into 18 ranges of adjusted gross income for the filer.

This chart lists three types of tax returns -- filers who have income for a child who earns more than $1,900 (meaning the child’s income is taxed at the parent’s rate); those who have income reported on Schedule D (primarily capital gains); and those without either of these types of income. For the purposes of our calculations, we are combining data for all three types of returns.
It's hard to see why PolitiFact went to all that trouble when the Congressional Budget Office--ordinarily a trusted source--has already done the work for them

PolitiFact crunched the IRS numbers, after a fashion, and didn't find much to vindicate Obama:
By these calculations, Obama would be incorrect in most cases.
I saw nothing in the calculations, inadequate though they might be, to indicate that Obama would ever be correct.  If any reader sees it differently I'd love to see the explanation.

Never fear, fans of liberal media bias.  There's more to come.

But that’s not the end of the story. These figures are for federal income taxes only. There are also a bunch of other federal taxes that could, and probably should, be included in the calculation. The burden for some some taxes, including corporate taxes, excise taxes and estate taxes, are hard to attribute to individual returns, so we’ll set those aside. But one federal tax is straightforward to throw into our calculations: payroll taxes.
Figuring the burden for corporate taxes, excise taxes and estate taxes may well provide a stiff test for researchers, but given the admitted relevance of those taxes why not make use of the previously-mentioned CBO report that estimates the effective tax rate with corporate and excise tax burden estimates figured in?

(Clipped from CBO report, click image for enlarged view)

Apparently PolitiFact's version of fact checking only involves consideration of the most regressive tax (by far) in the group, payroll taxes.  Ironically, payroll taxes are probably the least relevant tax in the group since the Social Security tax is peddled as retirement insurance--a premium paid for a fixed benefit package at retirement.  The argument for progressive insurance premiums based on the ability to pay lacks something in terms of moral authority.  Shall the rich also pay more per unit for milk, tea and gasoline?

We asked two researchers at the Urban Institute-Brookings Institute Tax Policy Center, Roberton Williams and Rachel Johnson, for their advice on how to factor in payroll taxes. They estimated that combining the workers’ share of the payroll tax with the employer’s share -- the usual practice among economists -- would mean an extra 15 percentage points for our hypothetical middle-class worker, and less than 2 additional percentage points for the high-income taxpayer.
Don't you just love the back-of-the-envelope methodology?

Let's take a figure calculated by journalists based on data that ignore a number of relevant taxes such as excise and corporate taxes, then call on the left-leaning Tax Policy Center to give us a modification based on (regressive) payroll taxes.  While we're at it, let's ignore the work done by the highly respected CBO touching the issue.

PolitiFact doesn't tell us what method (see Update below) the Tax Policy Center experts used to estimate an additional 15 percentage points for the middle-income worker.  Based on this chart from the Social Security Administration I'd hazard the guess that they assumed that the worker is self-employed.  Either that or they figured the employer's share of the payroll tax as a tax on the employee's income.

(Social Security Administration chart; click image for enlarged view)

For what it's worth, the CBO figured the effective tax rate for payroll taxes at 9.5 percent for both 2004 and 2005.  That figure seems very plausible if the CBO took into account both employees and self-employed workers, falling as it does between the 7.6 percent and 15.3 percent figures from the chart above.  The 9.5 percent figure matches well the expectation that most workers in the middle quintile are employees rather than self-employed.

It's kind of hard to take PolitiFact seriously on this fact check, but let's keep trying.

Adding these to the percentages we previously found for the income tax alone produces a new, "final" rate of 22 to 23 percent for the construction worker and 20 to 30 percent for the $50 million earner.

So it’s certainly possible that a given construction worker pays a higher effective tax rate than a given $50 million earner -- but it’s also not a guarantee. There is a lot of variation for taxpayers in both categories -- especially when you consider that we have been using average tax rates rather than median tax rates, which means that a small number of very high incomes can throw off the average.
Summary:  If you fudge the numbers enough in Obama's favor you can make it seem possible that he's possibly correct in some individual cases.

Just don't take the highly progressive effects of corporate taxation into account and you can believe that there's some basis in truth for Obama's claim.

Whoopee!  It's conclusion time:
Our ruling

Obama has made a sweeping statement that it’s common, even typical, for a $50,000-a-year construction worker to pay a higher tax rate than someone earning $50 million a year. The reality is that it’s hard to know for sure.

We found that a typical taxpayer with $50,000 in income pays 22 to 23 percent in income and payroll taxes, while the comparable rates for very, very wealthy taxpayers are in the 20 to 30 percent range. The data isn’t specific enough for us to be able to say if a majority of $50 million earners pay tax rates at the low or high end of that 20 to 30 percent range. We do think it’s safe to say that some of those very high earners pay a smaller share of their income in taxes than workers who earn $50,000, and some don’t. So, on balance, we rate the statement Half True.
The evidence supporting PolitiFact's view that "some high earners pay a smaller share of their income in taxes than workers who earn $50,000" comes from an extremely dubious process that ignored taxes that place a greater burden on the rich along with a very suspicious estimate of the effects of payroll taxes on middle-income earners.  The CBO's numbers seem far more reasonable on the face of it, and they do not support Obama's statement at all.

And one other thing.  PolitiFact takes context into account, right?  There's nothing in the story about the things the president claimed as the cause of the unfair application of the tax code ("all these loopholes and tax breaks").  The tax rate on capital gains is not a loophole.  It is an intentional aspect of the tax code designed to promote risk and investment.  Neither is it a "tax break" any more than the lower income person in the lowest tax bracket is receiving a "tax break" for not having his income taxed according to the rate in the next bracket up.

PolitiFact quotes the president's demagoguery but fails to call him on it.  Instead, we get the next best thing to partisan spin if it isn't 100% partisan spin.

The grades:

Louis Jacobson:  F
Bill Adair:  F

I'll assign "F" grades for relatively modest journalistic flaw given the importance I place on accuracy and objectivity where a story receives the label "fact check."   But this story is a stinker through and through, receiving the special tag "journalists reporting badly."  Any journalist not predisposed toward Obama's position ought to question a 15 percent markup for the effective tax rate based on payroll taxes.  It doesn't add up.


I'm tempted to hypothesize that PolitiFact's continued difficulty with providing accurate fact checks on these tax issues has to do with the Democrats' apparent reliance on the class warfare strategy for the coming election.  Rather than based on sympathy with the Democratic Party itself, journalists may feel sympathy with the class warfare angle.  It jibes with the way many journalists view their role in society, I suspect.

After Afters:

Some readers may find it fascinating to compare my critique of this PolitiFact item with that of an enthusiastic Kossack (language warning).

Update (Oct. 3, 2011):

I should have credited PolitiFact with a partial explanation for its procedure of adding 15 percentage points to the effective tax on the $50,000 income worker.  Yes, one can consider the employer's share of the OASDI payroll deduction as part of the employee's income.  But one does not then calculate the percentage tax paid using $50,000 as the denominator as PolitiFact ends up doing by simply adding the employer's matching 6.2 percent to percentage tax paid by the employee.  It is explained in slightly more detail here.

Update II (Oct. 4, 2011)

Turned the "m" right side up to make "may well" instead of its predecessor, "way well," and eliminated a stray two-letter word in the same paragraph.  Hat tip to Jeff Dyberg for pointing out the errors.

Wednesday, September 28, 2011

PolitiFlub: Engaging Rick Perry

I can't pass this one up, so it's time for another PolitiFlub item, the Cliff's Notes version of the Grading PolitiFact category.

A PolitiFact item stemming from Florida's iteration of the GOP presidential debate focused on this (bold emphasis added) statement from Gov. Perry in response to fellow candidate Rick Santorum (bold emphasis carried over from the version published by PolitiFact):
WALLACE: Governor Perry, 30 seconds to respond, sir.

PERRY: I've got one question for him.

Have you ever even been to the border with Mexico?


PERRY: I'm surprised if you have, but you weren't paying attention, because the idea that you --

SANTORUM: Well, the answer is, yes, I have.

PERRY: -- are going to build a wall, a fence for 1,200 miles, and then go 800 miles more to Tijuana, does not make sense. You put the boots on the ground.

We know how to make this work. You put the boots on the ground. You put the aviation assets --

SANTORUM: But it's not working, Governor.

PERRY: -- in the ground. No, it's not working because the federal government has not --

SANTORUM: But you said we know how it works. Is it working in Texas?

PERRY: The federal government has not engaged in this at all. When I'm the president of the United States, I'll promise you one thing --

SANTORUM: But you're saying you put the assets there. Has it worked in Texas?

PERRY: -- we will put the assets on the ground --

SANTORUM: You said you have.

PERRY: -- the boots on the ground --

BAIER: Senator Santorum, let him finish, please.

PERRY: -- the aviation assets on the ground, and we will stop illegal immigration, we will stop the drug cartels, and we will make America secure.
Hmm.  In what has the federal government not engaged?

Really? There aren't any federal boots on the Texas border?
Really?  PolitiFact seriously thinks Perry was saying the federal government hasn't put any agents along the Texas border?

If they're going to take it that literally then they might as well take Perry to be suggesting that we can secure the border by dropping boots from the sky to land along the southern border.  Boots as in shoes.

Basic charitable interpretation suggests that Perry was talking about something more than simply having agents on the border.  Even PolitiFact grants that it matters what the agents do while they're stationed there:
We're not checking whether federal investment has been effective — merely, as Perry said, whether the federal government has been "engaged."
This fact check item is a bad joke.  One cannot assume that Perry was denying the presence of federal agents on the border with his statement.  Perry's words carry a good deal of ambiguity stemming from his failure to specify what it is the boots ought to do while on the ground.  One can fault Perry for that, but it's simply ridiculous to leap to the assumption that he was denying the presence of federal agents.

The term "engaged" carries a connotation of focus.  As well, Perry may not have gotten around to describing the particular booted activity in which the federal government has failed to engage.  Moderator Chris Wallace terminated the exchange and Perry had only received permission for a 30 second response in the first place.

The PolitiFact item contains no hint that the PolitiFact team tried to obtain any clarification from Perry regarding his intent.  Therefore, the tag "journalists reporting badly" applies.


Tampa Bay Rays note

I made a couple of predictions related to the Tampa Bay (Devil) Rays this year. 

One was a specific prediction:  Jeremy Hellickson would win more games than the pitcher he replaced in the rotation, Matt Garza.  Garza had a decent season for the Chicago Cubs.  But Hellickson pitched very well all season to finish with 13 wins to Garza's nine.  W00t.  Nothing against Garza.  I just thought Hellickson's superior composure would give him the edge over Garza in earning positive decisions.  Both are good pitchers and I wish Garza well with the Cubs.

The second prediction was hazy, to borrow a term from the Magic 8-Ball.  I thought the Rays could finish anywhere from first to last, and that was a revised assessment after the Rays righted themselves after a dismal start.  At the start of the season I didn't think the Rays could win the division since I felt every team had improved except for the Rays.

But here we are.  With one game left in the regular season, the Rays stand tied with the Bo$ton Red $ox for the American League wild card playoff spot.

The credit goes to the pitching staff and the stellar defense, because the offense has struggled to score runs all season.

A few quick notes on the rest of the division:

Baltimore Orioles
I thought the Orioles would play much better this year.  But while the Orioles vastly improved on last year's season series performance against the Rays, Baltimore's young pitching staff just hasn't blossomed as did the Rays' over the past few years.  If the pitching comes along then watch out for Baltimore.  On the other hand, perhaps Baltimore's failure with its young pitchers serves as a tribute to Tampa Bay's organization.

Toronto Blue Jays
Toronto also starts a stable of young pitchers, and a number of them turned in solid seasons while the Jays' batting order hit dingers at a dizzying clip.  The Jays will finish near .500 in the tough AL East.

Boston Red Sox
Boston was the best team in the majors during the middle third of the season.  But the team couldn't play worth beans at the start and end of the regular season.  I accept no excuses.  Buying veteran players earns the risk of injury, and Boston had its share of those.  But the lineup was still better than it played near the start and finish of the regular season.  The Rays had no realistic shot at the playoffs without Boston's collapse.

New York Yankees
New Yorks batting lineup made its starting pitching stand up, with a good bit of help from a solid bullpen.  And the Yankees improved their play against the Rays this season.  Curtis Granderson had an insane season at the plate.  They earned the AL East title even if they paid a high price for it.

Monday, September 26, 2011

Grading PolitiFact: Barbara Boxer and the freeloading top 400

To assess the truth for a numbers claim, the biggest factor is the underlying message.
--PolitiFact editor Bill Adair

The issue:

(clipped from

The fact checkers:

Louis Jacobson:  writer, researcher
Martha Hamilton:  editor


PolitiFact's founding editor, Bill Adair, must fairly glow with pride when one of his PolitiFact teams cuts through the malarkey to identify the underlying point for a numbers claim.  The team of Louis Jacobson and Martha Hamilton pegged Sen. Boxer's underlying point in their opening paragraph:
During a Sept. 20, 2011, interview with Al Sharpton on MSNBC, Sen. Barbara Boxer, D-Calif., sought to highlight how America’s richest taxpayers have benefited economically in recent years.
PolitiFact provides all the context we're going to get without a subscription to Congressional Quarterly:
"The top 400 earners in this country are worth more … than half of the American people," Boxer said. "And since 1995, the top 400 wealthiest families have seen their incomes go up 400 percent and their tax rates go down 40 percent."

We checked the first part -- that the top 400 tax payers saw their incomes grow by 400 percent -- in March 2010, giving it a rating of True. But the second half of her statement was new to us, so we decided to look into it.
Apparently somebody watches MSNBC.  I'd like to know the full story behind the choice to grade this statement from Boxer.

When Boxer made her comments, the question of how much the wealthy should pay in taxes was at the top of the national agenda. In remarks on Sept. 19, 2011 about his plan for reducing the federal debt, President Barack Obama said that "middle-class families shouldn’t pay higher taxes than millionaires and billionaires. That’s pretty straightforward. It’s hard to argue against that. Warren Buffett’s secretary shouldn’t pay a higher tax rate than Warren Buffett." (Buffett is the legendary investor who for years has placed high on the list of richest Americans.)
The premise of Obama's statement is very probably false, though PolitiFact apparently can't be bothered to question it here.  The rich pay more in taxes.  They pay the same or higher tax rates, with the caveat that some types of income are taxed at a lower rate and some of the rich make a high percentage of their income from capital gains and investments.  But a middle-class taxpayer pays the same rate on those types of investments.  And in terms of the effective overall federal tax rate the rich again pay more than those in the middle class.

But at least PolitiFact continues to sniff around Boxer's underlying point.

As we explored elsewhere, Obama is correct that it’s possible for a secretary to pay a higher tax rate than a very wealthy person, but it’s also not typical.
PolitiFact's analysis failed to take effective federal tax rates into account despite the availability of a CBO report making clear the higher rates paid by higher earners.  The story tabled the concept of double taxation on income derived from corporate profits, which the federal government already taxes separately.  It may yet be possible for a secretary to pay a higher effective tax rate if the secretary spends all disposable income on gasoline and cigarettes--both heavily taxed by the federal government--but clearly that wasn't Obama's point.

In general, wealthier taxpayers do in fact pay a higher percentage in taxes than less-affluent people. The main exceptions are people in Buffett’s category -- the richest of the rich, whose income comes mainly from capital gains and dividends, which are taxed at 15 percent rather than the maximum of 35 percent for wages and salaries, and hedge fund managers, who benefit from a tax code provision that also taxes their earnings at 15 percent.
Note the continued failure through this point to examine the picture in terms of effective tax rates.  PolitiFact allows that most high income earners pay a higher tax rate while at the same time alleging a certain class of exceptions.  But PolitiFact provides flawed evidence in favor of the existence of those exceptions and grants itself, President Obama and Warren Buffett an exception to its "burden of proof" criterion.

In her comment, Boxer explicitly focused on Buffett’s peers.

Her data comes from a periodic Internal Revenue Service report that looks at the income and tax data from the 400 tax returns with the highest adjusted gross income. (In case you’re wondering, these 400 titans aren’t named, and their data is aggregated.) Technically, these aren’t "families" necessarily, but for a comment made during a television interview, we think it’s close enough.
The truth-finding experts at PolitiFact appear to misunderstand the data in the IRS report.  The top 400 earners are only "Buffett's peers" if Buffett himself appears on the list, and even then only in a sense perhaps too loose to float Boxer's point:
Over the 17 tax years a total of 6,800 returns were identified for the table. There were 3,672 different taxpayers representing the top 400 returns of each year. Of these taxpayers, a little more than 27 percent appear more than once and slightly more than 15 percent appear more than twice (see columns 2 and 3). In any given year, on average, about 39 percent of the returns were filed by taxpayers that are not in any of the other 16 years (see columns 4 and 5). In each year, 4 (or 1.0 percent) of the returns are for taxpayers who can be found in all 17 years. Thus, the data shown in the table mostly represent a changing group of taxpayers over time, rather than a fixed group of taxpayers.
At least four filers, and probably no more than that, occur in the top 400 for the first year of data collected as well as for the the most recent year.  To the extent that Boxer's statement implies a particular group of people receiving the benefit of a lower tax rate, that impression is misleading.  PolitiFact takes no notice.

We’ll take Boxer’s two claims in order. The data we’re using, which goes back to 1992, is adjusted for inflation.

"Since 1995, the top 400 wealthiest families have seen their incomes go up 400 percent."

In 1995, the aggregate adjusted gross income for all 400 tax returns worked out to $17.4 billion, or an average of $43.6 million per return.

By 2008 -- the most recent year available -- the aggregate income for the top 400 soared to nearly $66 billion, or more than $164 million in adjusted gross income per return.
It actually took me a bit of time to figure out what PolitiFact was doing with the above.  The PolitiFact team uses the column from the report giving all figures in 1990 dollars.  The method succeeds in adjusting proportionally for inflation, but offers a distorted picture of what the top 400 were making in 2008.  Aggregate income for the top 400 exceeded $108 billion in 2008 rather than the $66 billion figure from the story.

That amounts to a 276 percent increase. If Boxer had been using data one year older, she would have been correct: Using 2007 data, the rise was almost exactly 400 percent. This illustrates how sensitive this measurement is to economic conditions.

Credit to PolitiFact for doing the math correctly.  Boxer's figure ends up inflating the actual figure by nearly 45 percent.  PolitiFact probably stresses too little the sensitivity of the measurement.  Boxer, after all, is trying to make the case that the richest taxpayers are getting an increasingly good deal in recent years.  Certainly the trend for income is up, but it's a rocky trend.

(, click image for enlarged view)

Here's what it looks like graphically:

The first peak comes from tax year 2000 at the tail end of the Clinton economic expansion.  The second peak occurs in tax year 2007, at the end of the Bush economic expansion.  Boxer could have chosen to compare 1995 with the tax year 2000 and ended up with an increase of over 200 percent.  This helps illustrate the degree to which Boxer is cherry picking her data.  Additionally, one might detect a correlation between increases in income for the highest earners and good economic times.

None of this particularly assists Boxer's underlying point, needless to say.

"Since 1995 the top 400 wealthiest families have seen … their tax rates go down 40 percent."

From the same IRS document, we can see that the average tax rate for the top 400 returns was 29.93 percent in 1995 and 18.11 percent in 2008. That’s a decline of 39 percent -- for our purposes, close enough to Boxer’s claim to be accurate.
"Close enough" again involves ignoring the effective federal tax rate in favor of the average rate of income tax paid on adjusted gross income.  And inside the numbers we can see that between 1995 and 2000 the decrease in the average rate was 25.5 percent.  From 2000 to 2008 the decrease in the average rate was 19 percent.

Does that support Boxer's underlying point?  If not, perhaps PolitiFact should keep it quiet.

Note again that the this is not primarily a statutory change in tax rates.  Clinton did decrease the capital gains tax and George W. Bush did institute a broad tax decrease that affected the tax rates of the top 400 earners.  But the average tax measurement varies primarily because the top earners increasingly receive a greater percentage of income from capital gains.  So, contrary to what Obama, Boxer and PolitiFact assert, the question isn't really what percentage of their earnings the wealthy ought to pay, but rather whether it is a good idea economically to tax capital gains at a rate roughly equal to the second lowest marginal rate on salary income.

Admittedly, putting it that way takes the fun demagoguery mostly out of the equation.

We near PolitiFact's conclusion:
On the data, Boxer is right for one claim and wrong for the other, though on the one for which she’s incorrect, she would have been right using data from one year earlier.

More broadly, even the smaller, 276 percent increase supports her underlying point -- that for the richest of the rich, incomes have grown by leaps and bounds while average tax rates have declined for the same group by 40 percent.
More precisely, Boxer's numbers are correct for one claim but not the other, and neither set of numbers supports the underlying point that "America’s richest taxpayers have benefited economically in recent years" from tax policy except under a fairly loose concept of "benefit"--earning a higher percentage of their overall income from capital gains and investments.  That seems like a dubious measure of benefit.  The statutory rates haven't changed all that much.
As we’ve noted, extrapolating tax-burden patterns from the super-rich (like these 400 tax-paying units) to the merely very rich, the rich or the plain old affluent should be done with care. As the Congressional Budget Office has determined, the tax system is generally progressive, despite the lower tax rates for Buffett and his peers. The poorest one-fifth has an effective tax rate of 4.3 percent, the next fifth has a rate of 10.2 percent, the middle fifth has a rate of 14.2 percent, the second-highest has a rate of 17.6 percent, and the top fifth has a rate of 25.8 percent. And the rates keep going up the higher you climb. The top 10 percent of earners had an effective tax rate of 27.5 percent, the top 5 percent had a rate of 29 percent, and the top 1 percent had a rate of 31.2 percent.

In other words, the top 400 have a distinctly different tax pattern than even the richest 1 percent of American taxpayers, a much larger group that includes about 1.1 million households with an average pre-tax income of $1.7 million.
Ladies and gentlemen, we have a howler.

The PolitiFact team distorts or misunderstands the import of the CBO study it finally got around to citing.  The CBO report refers to the effective tax rate, which includes the estimated impact of corporate taxation.  Apply the impact of corporate taxation to the richest 400 earners, which the IRS report doesn't touch, and it will probably boost the effective tax rate up very close to the rate paid by the rest of the top 1 percent because corporate taxes have a greater impact on investment and capital gains income.  It's the old double taxation issue.

Compare PolitiFact's attitude on this fact check to the one exhibited in a fact check of Michele Bachmann from not too long ago:
Bachmann would have been right if she’d said, "the top 1 percent of income earners pay about 40 percent of all income taxes into the federal government." But she didn’t say that -- and even if she had, her decision to focus on income taxes, rather than looking at the whole federal tax picture, would have presented the numbers in such a way that wealthier Americans would look more heavily taxed than they are.
Likewise, Boxer and Obama find it more favorable to talk about income tax "rates" without mentioning that they are averages and while ignoring data on effective tax rates.  PolitiFact did the same thing before finally taking notice of the CBO report on effective tax rates--and then reporting on it so inaccurately that it might as well have been ignored.

The grades:

Louis Jacobson:  F
Martha Hamilton:  F

Why does the PolitiFact team give Boxer a pass on her stinker of an underlying point?  Probably because they believed that point with relative confidence prior to doing the fact check.  That's how ideological bias works when it manifests unintentionally.  The troublesome facts receive a faulty interpretation that helps them jibe with the accepted view.

Sunday, September 25, 2011

Grading PolitiFact: Nancy Pelosi and The Great Job Creator

When we find we've made a mistake, we correct the mistake.
--Principles of PolitiFact and the Truth-O-Meter

The issue:

(Clipped from

The fact checkers:

Louis Jacobson: writer, researcher
Martha Hamilton: editor


Back in July of this year I published the observation that PolitiFact uses wildly inconsistent criteria when evaluating job creation claims.  PolitiFact sometimes finds it important to distinguish between job creation and net increases in the number of jobs:
Our research took us down a road traveled by PolitiFact Texas in examining various claims from Perry that most of the jobs created in the U.S. are in the Lone Star State.

PolitiFact Texas has debunked several of these eye-popping claims, noting that they are based on statistics from a minority of states that have had overall job gains in recent years.
I likewise noted how the inconsistency tended to benefit Democrats while harming Republicans.

Is the distinction important to PolitiFact in this case involving Pelosi?

Let's find out, taking it from the top:
During a Sept. 17, 2011, edition of the MSNBC show Up with Chris Hayes, House Minority Leader Nancy Pelosi, D-Calif., offered a defense of the U.S. economy’s job creation record during President Barack Obama’s tenure.

"When President (George W.) Bush was president for those eight years with those tax policies, we lost jobs," Pelosi said. "In fact, more jobs were created in the second year -- that would be last year -- in the private-sector in the Obama administration than in the eight years of the Bush administration. And they want us to repeat those tax cuts at the high end which, again, did not create jobs, but they did deepen the deficit."
Pelosi makes no apparent attempt to distinguish between net job creation and gross job creation, though in fairness to her it is just as clear what she's talking about in context as it is in the other cases PolitiFact evaluated.

Unfortunately for Pelosi, it's also pretty obvious that she's blaming the job losses near the end of Bush's second term on tax policies implemented during his first term--a classic case of misleading with cherry-picked stats.  To accept her reasoning we would have to rule out the deflation of the housing bubble and the associated problems in the derivatives market as causes for those job losses.

PolitiFact calls Pelosi on the cherry-picking yet somehow succeeds in making it appear that Pelosi makes a good point (bold emphases added):
Pelosi’s statement from last weekend continues the cherry-picking approach: She cited numbers only for jobs created in 2010.

Using the period January 2010 to December 2010, the number of private-sector jobs increased by more than 1.2 million. If Pelosi had used more current numbers, she actually would have had a stronger case: From January 2010 to August 2011, the increase in jobs was almost twice as big -- nearly 2.4 million.

Even the smaller of these two numbers beats Bush’s eight-year jobs increase -- but she continues to conveniently ignore the first year of the Obama administration. If you take the entirety of Obama’s term -- January 2009 to August 2011 -- the nation has lost 1.8 million jobs, which is still worse than what happened under Bush.

Nadeam Elshami, a spokesman for Pelosi, argues that it’s proper to keep the first year of Obama’s tenure out of the equation.
Elshami's rationale is priceless owing to its bankruptcy:
"Republicans may wish to wipe the slate clean and forget about the financial crisis that nearly destroyed our economy and the recession that began under President Bush," said Elshami, a spokesman for Pelosi. "The figures are correct and clearly demonstrate that private-sector jobs are being created, but not fast enough – we have much more to do, and we urge Republicans to join us to pass President Obama’s American Jobs Act."
Far from wanting to forget about the financial crisis, Republicans would want to point out its existence in order to undermine Pelosi's apples-to-oranges comparison.

But again we find PolitiFact offering a torqued corrective (bold emphasis added):
We understand the impulse to make such adjustments, but in Pelosi’s formulation, this principle is applied unequally. If the first year of the Bush administration, which also included a recession, were excluded from her calculation-- this would have magically erased a loss of 2.4 million jobs under Bush. The next seven years of the Bush presidency would have shown a gain of nearly 1.8 million jobs, rather than a loss of 653,000 for the full eight-year period.

Interestingly, if Pelosi had started the count for each president at the beginning of the second year of their presidencies, it would not only have been more defensible methodologically, but it would have also backed up Pelosi's argument better. Leaving out the first year for each president, private-sector job growth under Obama would have been 2.4 million, compared to 1.8 million under Bush.

However, in none of the three statements by Pelosi that we've fact-checked did Pelosi use that method.
If we simply cut out the first year for each president then Bush is relieved of the effects of one recession on job creation while leaving the effects of a second recession.  Obama simply has recessions cut cleanly away.  And that would support Pelosi's argument?  Nonsense.  It leaves us with more apples and oranges.

Pelosi compared a select time frame in the Obama administration against the entire length of the Bush administration -- a methodology that treats the two presidents unequally. The irony is that if she had used better methodology, she would have had a sounder argument that more private-sector jobs were created under Obama than under the Bush administration. For her general point, we give Pelosi some credit. For her methodological sins -- repeated at least three times -- we give her thumbs down. On balance, we rate her statement Half True.
In the conclusion we again see PolitiFact trying to push Pelosi's underlying argument even while they criticize her details:  "For her general point, we give Pelosi some credit."  That point doesn't fly even with PolitiFact's spin added.

Because of Pelosi's "methodological sins" PolitiFact drops her down to "Half True"--yet PolitiFact Virginia gave Gov. Bob McDonnell of Virginia a "Barely True" grade for a claim that was otherwise accurate by the numbers and in terms of its underlying argument simply for his failure to distinguish between gross job creation and net job creation:
But the Virginia governor used net job numbers. Those are different from job creation figures, which are not compiled by the BLS or other national authority. So using net job numbers and calling them job creation figures is a case of comparing apples to oranges.

We rate McDonnell’s claim Barely True.
PolitiFact's rating of Pelosi stands as incredibly generous considering that her point comparing the tax policy of Bush with that of Obama finds absolutely no support in either her version or the PolitiFact spin version.

The grades:

Louis Jacobson:  F
Martha Hamilton:  F

Pelosi's methodological sins barely eclipsed those of the PolitiFact team, and Pelosi didn't have the temerity to call her statement a fact check.


Here's a little update on the winners and losers with respect to PolitiFact's favoritism:

Helped by omission
 Rick Perry (True)
 Terry McAuliffe (True)
 Carolyn Maloney (True)
 Brian Moran (True)
 Mitt Romney (True)
 Austan Goolsbee (Mostly True)
 Ed Gillespie (Mostly True)
 Joe Biden (Mostly True)
 Nancy Pelosi (Half True)
 Nancy Pelosi (Half True)
 Debbie Wasserman Schultz (Half True)
 Rick Scott (Half True)
 Tim Pawlenty (Pants on Fire; harmed by omission?)
 John Boehner (False; harmed by omission)

Harmed by inclusion:
 Texas Public Policy Foundation (Half True)
 Rod Smith (False)
 Rick Perry (False)
 Bob McDonnell (Barely True)
 Rick Scott (Pants on Fire)

Correction:  Had incorrect links for the Rick Perry ratings in two spots.  Hat tip to "KnocksvilleE" for pointing out the problem.

Wednesday, September 21, 2011

Grading PolitiFact (Wisconsin): Al Sharpton and the draconian milk cuts (Updated x2)

We always try to get the original statement in its full context rather than an edited form that appeared in news stories.
--About PolitiFact

The issue:

(clipped from

The fact checkers:

Tom Kertscher:  writer, researcher
Greg Borowski:  editor


MSNBC host Rev. Al Sharpton made his comments on MSNBC's "Politics Nation," aptly enough.  The transcript (yellow highlights indicate the portion quoted in the PolitiFact story):
SHARPTON: Republican politicians throw around budget numbers like  it`s all some abstract game. But they seem to forget their policies have a devastating impact on real people across the country. In Michigan,  Governor Rick Snyder has signed the law that puts a four-year lifetime limit on welfare benefits. That means next month, nearly 41,000 Michigan residents will stop getting welfare aid including nearly 30,000 children. Thousands of innocent impoverished kids, abandoned by the state. In Arizona, a new republican law punishes the families of prison inmates. To help deal with the state`s deficit, the state has begun charging $25 to visit someone in prison. It`s called a quote, "background check fee." But actually goes for maintenance and repairs.

And in Wisconsin, Governor Scott Walker`s budget cuts means, some kids are going without. One school district is so worried about losing state funding, that has stopped giving milk to elementary school kids during snack time. These are the real government of GOP priorities.
 Aside from some journalistic impropriety, the quotation presents Sharpton's statement reasonably well.

I'll take a moment to pan the journalists for the presentation of the quotation.  PolitiFact lists the MSNBC transcript as the source.  The transcript differs from the version of Sharpton's words used by the writer, Tom Kertscher, and passed on to the reader by the editor, Greg Borowski.    The PolitiFact Wisconsin version cleans up some faulty grammar, including a missing "it," in the MSNBC transcript.  Sharpton may or may not have made his statement the way PolitiFact presented it, but the transcript isn't the source of the quotation if the two differ unless PolitiFact Wisconsin relaxed its stance on the normal journalistic standard for handling quotations.

On with the fact check:

Friday, September 02, 2011

Grading PolitiFact: Robert Reich, corporate profits and "just before" the Great Depression

Words matter -- We pay close attention to the specific wording of a claim. Is it a precise statement? Does it contain mitigating words or phrases?
--Principles of PolitiFact and the Truth-O-Meter

The issue:

(clipped from

The fact checkers:

Louis Jacobson:  writer, researcher
Martha Hamilton:  editor


PolitiFact makes a show of insisting that words are important.  "We play close attention to the specific wording of a claim," reads a portion of PolitiFact's page of fact checking principles.

Unfortunately, PolitiFact's application of this principle, as is the case with many of its other principles, ends up a matter of convenience.  Sometimes the specific wording of a claim matters but sometimes it doesn't.  And good luck discerning the criteria for the varying application of the standard.  This fact check of liberal economist Robert Reich serves as the latest example.

Reich's statement occurs as its own paragraph in the midst of an op-ed published at a (no doubt non-partisan) PBS site. Reich builds a comparison between the modern state of American labor and the state of labor during the Great Depression, comparing wage growth between the two periods, for example.  Reich's statement, in context, extends that comparison:
The ratio of corporate profits to wages is now higher than at any time since just before the Great Depression.

Meanwhile, the American economy has all but stopped growing – in large part because consumers (whose spending is 70 percent of GDP) are also workers whose jobs and wages are under assault.

Perhaps there would still be something to celebrate on Labor Day if government was coming to the rescue. But Washington is paralyzed, the president seems unwilling or unable to take on labor-bashing Republicans, and several Republican governors are mounting direct assaults on organized labor (see Indiana, Ohio, Maine and Wisconsin, for example).
The PolitiFact story reveals that Reich's statement came under review at the request of a reader.

On with that review:
We turned to statistics compiled by the Bureau of Economic Analysis, the federal office that calculates official statistics about the economy. We found numbers for corporate profits as well as for two measures of worker income -- wage and salary disbursements, and total employee compensation received. We then divided corporate profits by both of the income measurements, all the way back to 1929. (Here are the full statistics from 1929 to 2011 as we calculated them.)

Do click the link and review PolitiFact's numbers.

For wages, we found that Reich was essentially correct. The ratio in 2010 -- the last full year in the statistics -- was .281, which was higher than any year back to at least 1929, the earliest year in the BEA database. The next highest ratio was in 2006, at .265. (We didn’t find pre-1929 data, so the one part of Reich’s statement that we can’t prove is that the ratio was higher "just before the Great Depression.")
Excuse me, but the "just before the Great Depression" part constitutes a key part of Reich's overall argument. His op-ed, as mentioned, builds a comparison between modern times and the Great Depression and tries to pin the blame for the similarity on business and government inaction.  Take away the reference to the Great Depression and the statement no longer suits Reich's purposes.  We end up with a modern ratio slightly higher, for example, than the one observed during World War II.  Reich doesn't want workers toiling loyally as though we're at war (even if we are at war).  He wants people to push for government action such as many people believe helped alleviate the Great Depression.

We have yet another case where PolitiFact pays close attention to the specific wording of the claim so that its agents can make sure to give a pass to liberals who say indefensible things.  Applying PolitiFact's (fallacious) "burden of proof" grading criterion would result in giving Reich a poor rating such as "False."  By ignoring the part of the claim for which Reich would have borne a burden of proof--changing the claim, in effect--PolitiFact can give him a better grade.

I can't wait to see how this one turns out.

We also looked at total compensation, since the portion of worker compensation delivered outside of wages has grown significantly since 1929. The numbers were slightly different, but the general pattern still held. The ratio in 2010 was .226, which was matched or exceeded in only four years -- 1941, 1942, 1943 and 1950.
Only four exceptions to Reich's claim using that method?  That certainly sounds like trouble for Reich even if he gets a pass on "just before."
To capture the most up-to-date trends, we also looked at the ratios for the last six quarters. For both wages and compensation, the ratio has risen steadily over that year-and-a-half period. For wages, the ratio has climbed from .274 in the first quarter of 2010 to .290 in the second quarter of 2011. For compensation, the ratio has risen from .220 in the first quarter of 2010 to .234 in the second quarter of 2011.
Trends?  Reich doesn't argue trends.  Trends are irrelevant to his argument.  He's asserting a problem now regardless of trends and trying to emphasize the severity of the problem by using the comparison to numbers preceding the Great Depression.

PolitiFact continues:
So numerically, there’s little question that Reich is essentially right. (Or, at least for now he is. Economists note that statistics about corporate profits and wages are often revised after the fact.) A more interesting question is what this trendline actually means.

Little question that "Reich is essentially right"?  PolitiFact isn't even checking what Reich said!  Reich gets graded on a new claim invented for him by PolitiFact:    Goodbye, "since just before the Great Depression."

PolitiFact's eagerness to rescue Reich's op-ed via the emasculation of one of his key points is all the more outrageous given yet another of PolitiFact's viscous principles:
Context matters -- We examine the claim in the full context, the comments made before and after it, the question that prompted it, and the point the person was trying to make.
PolitiFact couldn't verify Reich's claim and changed it to something verifiable that failed to help Reich make the point he was trying to make in his op-ed.

Welcome to the wonderful world of PolitiFact fact checking.

The fact check goes through a series of paragraphs explaining the numbers in a manner that further undercuts Reich's point.  Try this one, for example:
Typically, businesses initially lose ground during a recession, while workers suffer somewhat less, in part due to "sticky wages" -- the tendency for worker pay to increase or stagnate rather than fall, even in hard times. This pattern tends to decrease the ratio of corporate profits to wages.
If you're wondering how this meshes with Reich's claim of a high corporate profits to wages ratio just before the Great Depression then join the club.  But don't expect PolitiFact to explore the discrepancy. 

Let's skip to the end:
On the numbers, Reich’s claim is essentially correct. And in his analysis, Reich doesn’t over-promise on what the data indicate. Amid evidence that these numbers could turn out to be a temporary spike, he resists the temptation to label it the culmination of a long-term trend. We find Reich’s formulation both factually supportable and appropriately cautious in its interpretation. We give it a rating of True.
This is flippin' ridiculous.  "Reich doesn’t over-promise on what the data indicate."  Right, because PolitiFact didn't check the datum Reich was using to make his point, which was the supposed corresponding peak of the corporate profits to wages ratio at the time "just before the Great Depression."  Reich also carefully avoided confusing "effect" with "affect" by using neither term in this op-ed.  Sheesh.

PolitiFact went into full spin cycle for Reich.

The grades:

A quick review:
  • Ignore a big part of Reich's claim ("just before the Great Depression")
  • Ignore the context of the claim
  • Ignore the underlying message Reich was trying to make
  • Ignore four exemptions to Reich's claim under one of the methods PolitiFact used to test the claim
  • Ignore "burden of proof" criterion

Louis Jacobson:  F
Martha Hamilton:  F

I'd imagine shoddy fact checking like this can emerge consistently from an organization only if it is overpoweringly overconfident in its own even-handedness and competence.  It makes me think of a Southern lady who once drawled "Ah don't have an accent!"