A comeuppance for PolitiFact and Rex Nutting (Updated)

The math simultaneously backs up Nutting’s calculations and demolishes Romney’s contention. The only significant shortcoming of the graphic was that it failed to note that some of the restraint in spending was fueled by demands from congressional Republicans.
--Louis Jacobson, PolitiFact

(clipped from PolitiFact.com;
click image for larger view )

The White House, with an assist from PolitiFact, has turned Rex Nutting's analysis of President Obama's spending into a key talking point for Obama's re-election campaign.

While I'm neither an accountant nor an economist, I've put together enough evidence to show that Nutting and PolitiFact are selling falsehoods as the truth.

Cherry picking the metric

As I illustrated in an earlier post (at PolitiFact Bias), the "annual average" increase in spending serves as a very misleading method of measuring increased spending.

Cherry picking the window

Nutting compared presidents according to a four-year window of performance.  But Obama has only spent money over a span of less than three years (not counting FY 2009).  This allows Nutting to hide Obama's spending increases over a four-year average by figuring in the constraints of a Republican House of Representatives and the loss of a filibuster-proof Senate majority.

Failing to adequately contextualize Obama's actual spending

PolitiFact admitted in an update that Obama had an unusually active role in spending during FY 2009.  Obama and Congress passed an expensive economic stimulus package among other spending bills.  Nutting and PolitiFact dropped Bush's spending in FY 2009 by $140 billion to offset that spending, but that fails to account for bailout programs like TARP.  TARP involved great sums of money disbursed in unusual ways, so any fact check failing to nail down the major impacts of bailouts on spending for FY 2009 will not definitively address the issue of spending increases relative to the FY 2009 baseline.  Nutting certainly made no serious attempt, and PolitiFact shamefully overlooked that failure.

Cherry picking the future projections for the cherry-picked window

Cherry picking the window was bad enough, but Nutting and PolitiFact amplified the error by also cherry picking the future projections used to complete the picture of Obama's window.  The Office of Management and Budget made future projections.  Nutting did not use the OMB projections despite his use of OMB data for figuring the spending increases of earlier presidents.  Instead, Nutting (and PolitiFact) used the lower of two CBO projections.  Needless to say, the CBO projections used were less realistic as well as more favorable to Obama than the other two.

Shielding a fact check from scrutiny

PolitiFact obscured the reasoning behind its rating of Nutting's claim by simultaneously ruling on a statement from Mitt Romney.  PolitiFact did not answer a request that it explain how it accounted for $140 billion taken from Bush's 2009 spending baseline and instead supposedly credited to Obama.  Was that amount added to Obama's FY 2009 spending or did Obama obtain the benefit of a spending Twilight Zone in FY 2009?


Fixing the problems

The right metric, the right window

The best metric for comparison involves a figure that allows for an apples-to-apples comparison while avoiding the distortions inherent in using the average annual increase.  Using the percentage increase in spending as measured by the OMB for Obama's two complete years comes closest, though it leaves us the challenge of making appropriate adjustments to account for the unusual spending conditions in FY 2009.  The inflation-adjusted percentage increase for a two-year span--no averaging--provides the right window for assessing the claim from Mitt Romney. 

Contextualizing the spending

This part's tricky.  So far I don't have definitive answers as to how the government treats bailout spending.  But I have found a few things.  Economist Greg Mankiw, for example, wrote that the U.S. Treasury and the CBO used different methods to account for TARP spending, with the latter figuring TARP spending at about one-fourth its total to account for anticipated repayments.  For FY 2009, the Treasury Department put the TARP spending figure at $364 billion (over 10 percent of FY 2009 spending):

As of September 30, 2009, approximately $317 billion of the $700 billion in purchase and guarantee authority remained available. During FY 2009, Treasury-OFS disbursed $364 billion, most of it in the form of investments. A total of $73 billion of those TARP funds have already been repaid during FY 2009, resulting in a total of $291 billion in investments outstanding as of September 30, 2009 (see Chart 6).

The Treasury's statement for the following year indicated modest spending under TARP in FY 2010, in the amount of about $24 billion:

Since TARP’s inception on October 8, 2008 through September 30, 2010, Treasury has disbursed $387.7 billion in direct loans and investments under TARP. Over half ($204.1 billion) of those funds has been repaid, and the investments have generated $27.8 billion from cash received through interest and dividends, as well as proceeds from the sale and repurchase of assets in excess of cost. As of September 30, 2010, TARP had $179.2 billion in gross outstanding direct loans and equity investments, valued at $142.5 billion.

The method Nutting and PolitiFact used would apparently put all $364 billion in TARP spending on Bush for FY 2009 even though about half of the disbursements were directed by the Obama administration.

Some critics allege that TARP repayments come off of Obama's spending numbers as "negative spending."  While I am so far unable to absolutely confirm that claim, the AP's fact check on this issue appears to accept it.

The bottom line for now:  We don't have enough information to rule definitively on Romney's claim that Obama accelerated spending more than any president in recent history.  The AP fact check puts Obama's first two years in close competition with Reagan's biggest jumps in annual spending growth.

And those future projections ...

Just throw out the future projections for purposes of fact checking Romney's claim.  They aren't relevant in that context.  Romney talked about what Obama had already done, not what was projected for the future.  We can, if we wish, use the projections to make tentative comparisons between four years of Obama's spending with four years of spending from other presidents, though we end up with a difficulty in adjusting for future inflation.  The future estimates from the OMB serve much better than the unrealistic CBO baseline, and are somewhat kinder to Obama than the CBO's alternative scenario.   I use the OMB numbers for my calculations.

Putting it all together

I adjusted the FY 2009 baseline downward by $307 billion, which represents the lower of two adjustments PolitiFact mentioned other than Nutting's $140 billion.  A higher adjustment is justifiable but unnecessary to drive the point.

First, I adjusted all of the OMB figures and the modified FY 2009 figure for inflation.

Then I added each pair of consecutive years together to obtain an apples-to-apples basis for comparison, keeping the focus mostly on pairs led by but one president.  This method allows us to compare Obama's performance in 2010-2011 with every other pair of years from Reagan forward.  After combining total spending for each pair of years, I calculated the percentage increase over the preceding pair of years to obtain a two-year figure for the percentage increase in spending.

The steepest increase occurred during the pair shared by Bush and Obama, and second place went to Obama's 2010-2011.  A graph like Nutting's with his figures replaced with mine looks like this:

Click image for larger view

Nutting's work was hugely misleading and PolitiFact's fact check was a joke.  Obama's big spending was real, and PolitiFact ought to correct the record properly.  I'll repeat for the sake of clarity that my analysis uses numbers from the OMB and combines those with numbers PolitiFact told us did not affect the accuracy of their analysis (like the $307 billion adjustment and inflation adjustment).  The annualized average was a key misleading element of Nutting's work and PolitiFact never figured it out.  Either that or just decided to keep it quiet.

Other mainstream fact checks also appeared blind to this key element of Nutting's analysis.


I'll update after I expand my calculations back as far as President Eisenhower.


Update 6/17/2012:

Using the biennial metric for comparison with Obama's two-year spending window produced results far different from Nutting's (find the Google spreadsheet I used here; please bring any errors to my attention).

Nutting's claims about Obama's supposed low spending relied on his misleading technique of using annualized averages.  Using one-year increases in isolation, Obama's FY 2010 spending (adjusted for inflation) put him among the top five spending accelerators going back to President Carter.  Using that metric Nutting is wrong and Romney may be right depending on resolution of all the issues surrounding FY 2009.

The biennial metric made things worse for Obama.  Not counting FY 2009 and its unresolved issues, one has to go back to President Ford to find higher increases in spending percentage on a biennial basis.  PolitiFact considered Ford's tenure too short for comparison, so skipping Ford we'd have to go all the way back to L.B.J. to find higher spending than Obama's.  None of this jibes with a claim that Obama's spending increases are the lowest in 60 years.

Nutting and PolitiFact are clearly wrong.  Romney, in contrast, may be right.



June 16, 2012: Updated graphic with opaque background for improved larger view.  Also on June 16 (not June 6!) fixed a typo in the Bush 2 '03-'04 column.  The percentage change originally read as "10.02" percent.  The change is also noted in the text of the updated graphic.