Grading PolitiFact: President Obama and job loss timing

Does it take want-to in order for a journalist to ignore an inconvenient underlying argument?

The issue:

The fact checkers:

Martha Hamilton:  writer, researcher
Bill Adair:  editor


Analysis:

PolitiFacter Martha Hamilton wrote a fact check relating to President Obama's recent appearance on the Jon Stewart Show.  Note that the second quoted paragraph implicitly acknowledges an implicit, or underlying, argument from the president:

 "We lost 4 million jobs before I was sworn in; 750,000 the month I was sworn in; 600,000, the month after that; 600,000 the month after that. So most of the jobs that we lost were lost before the economic policies we put in place had any effect."

Candidates have sparred endlessly over who deserves the blame for the nation’s high unemployment rate -- 9.6 percent in the latest tally. After the Daily Show interview, one of our readers asked us to check Obama's numbers. So we decided to take a look.

Who deserves blame for the nation's high unemployment rate?  Do Obama's numbers have any particular bearing on that?  Hamilton doesn't hazard a guess.  But it must have a direct bearing if PolitiFact bothers with a fact check without explaining that the policy implication doesn't follow.  Right?

Beware, dear reader.

Though Hamilton skipped quickly past, Obama did present an underlying argument with his above statement.  That is, since most of the jobs lost since he took office occurred before his economic policies were put in place, therefore (probably) his economic policies were responsible for slowing the loss of employment.  But recessions are part of the business cycle, so purported explanations should receive careful examination and correlations examined closely before concluding causation.

PolitiFact will not be conducting any kind of examination of the correlations other than to determine whether they exist in the first place.  We'll follow Hamilton through the disturbingly brief process:

We found a match: Looking at BLS data on seasonally adjusted non-farm employment from December 2007, when the recession officially began, to January 2009, the month before the stimulus was enacted (a 25-month period), the jobs number declined by 4.4 million. So Obama’s first number was right, although he could have been clearer about the time frame.

This so-called match doesn't have much to do with the underlying argument noted above.  It has more to do with a different underlying argument (Bush's fault).  Since that number is relatively meaningless there's not much point in double-checking it.  It's not controversial that many jobs were lost at the tail end of the Bush administration, and those occurred owing to a complex set of causes.

What about the claims PolitiFact posted to the item description, like "Most of the jobs we lost were lost were lost before the economic policies we put in place went into effect"?  Based on the stats Obama mentioned, it makes sense to take the lost jobs Obama talks about as jobs lost after he took office and before his policies went into effect.

It's unclear whether Hamilton sees it the same way:

When he refers to his economic policies, we presume he is referring to his main economic stimulus, the American Recovery and Reinvestment Act. It passed in February 2009, but it took several months before the impact of its spending was felt in the economy.

Job loss didn’t stop, but Obama is right that it slowed down. In the 19 months from February 2009 through September 2010, the month of the most recent preliminary data, the overall job decline in the private and public sectors was 2.6 million. And the number of jobs lost per month has declined from around 700,000 a month at the beginning of the administration to months in which there were small net gains. Since May, however, the losses -- albeit smaller ones -- have returned, giving Republicans fresh ammunition. For example, payroll employment dropped 57,000 between July and August 2010.

Hamilton provides information relevant to the issue I identified, but does not appear to address Obama's claim.  Instead of comparing 4.4 million lost jobs attributed to Bush to the 2.6 million lost under Obama, she ought to have compared job losses under Obama prior to implementation of his economic policies to job losses after implementing those policies.

PolitiFact's numbers from the Bureau of Labor statistics showed "around 700,000" jobs lost in Jan. 2009.  Obama claimed 650,000.  I'm not sure of the exact source of either figure.  I constructed a graph using the BLS's monthly employment press releases (like this one).  The figure for Feb. 2009 was 651,000--very close to the figure Obama offered for January.  Was that his source?  I don't know.  Regardless, here's the graph (click to enlarge):

• The skinny vertical red line represents Obama taking office.
• The skinny vertical blue line represents passage of the economic stimulus bill.
• The transparent red block represents a period greater than two months but less than many (the definition of "several," following the reasoning offered above by Hamilton.  Feel free to imagine it stretching further to the right to match your expectation for less than "many."
• The Bureau of Labor Statistics credits over 400,000 of the April 2010 jobs to temporary work with the Census Bureau.  In short, the high spike is because of the Census.

Judging from the way Obama phrased his argument, he does not expect job losses in February and March to count against his record.  He's taking a minimalistic view of the window of "several" carved out for him by PolitiFact's Hamilton.

But there's a problem.  By taking credit for the job losses from March through April, the president puts himself on the hook for 1.9 million jobs lost (at least for the set of figures I'm using).  That compares to 1.95 million for the first three months according to the figures Obama used--barely "most," though pro-rating the numbers would give him a beneficial cushion.   And if Obama was to say that it took longer for his policies to take effect then they're going into effect after the job loss numbers have already begun to trend downward.  The latter doesn't lead to a strong inference toward his policies as the reason for the reversal of the trend.

Where does this leave us in terms of fact checking?  Hamilton may be right that Obama is claiming that most of the jobs lost during the recession occurred before his policies went into effect.  That makes for a very weak implicit underlying argument, so I think it charitable to believe that Obama is arguing that while on his watch most of the job losses occurred before his policies were implemented.  The figures I used total 1.91 million jobs lost for the first three months of his administration.  That's slightly less than half the total of 1.94 million lost since that time under Obama.  It's very close, then, and another set of figures might show the president correct--barely--and with a slightly improved but still weak inference in support of the effectiveness of his policies.  The literal statement is probably true and the underlying argument is dubious.

The grades:

Martha Hamilton:  F
Bill Adair:  F

I fault both PolitiFacters for ignoring the alternate interpretation of Obama's statement and the underlying argument.  Plus the citations were not nearly as helpful to readers as they might have been.

Afters:

I've been surprised at how difficult it is to locate graphic material on job loss trends.  I'll probably pretty up the data I've collected and start performing updates like I did for the Iraq War casualties.


Nov. 8, 2010:  Deleted "drop in" where it preceded "job losses" in the second paragraph of analysis following the graph.  The original wording created something like a double negative.
Nov. 29, 2010:  Corrected spelling of Jon Stewart's name, omitting the "h" from his first name.  Also removed a sentence fragment for which I could no longer recall the completed version (second paragraph following the graph).  Reworked the second paragraph following the graph to assist coherence.