Thursday, January 27, 2011

Grading PolitiFact: Talking points from the Van Hollen twins

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

The issues:

click to enlarge


The fact checkers:

Group A
Robert Farley:  writer, researcher
Louis Jacobson:  writer, researcher
Martha Hamilton:  editor

Group B
Robert Farley:  writer, researcher
Louis Jacobson:  writer, researcher
Martha Hamilton:  editor


Analysis:

Both fact checks come from the same Van Hollen utterance:
“I’m interested to hear my colleagues say that they can identify with all the problems in the health care system.  Between the year 2000 and 2006, premiums in this country doubled, health insurance company profits quadrupled, and this Congress did nothing."
(yellow and blue highlights added)
The yellow highlights indicate the portion of the quotation that PolitiFact chose to fact check.  The blue highlights indicate where I believe Van Hollen expressed his underlying point.

This is what Van Hollen looked like in action, for those interested:




Claim A

PolitiFact:
In fact, we first visited a very similar claim when Sen. Jay Rockefeller, D-W.Va., wrote a Sept. 21, 2009, opinion piece for the Capitol Hill newspaper Roll Call in which he said, "Insurance companies have seen their profits soar by more than 400 percent since 2001, while premiums for consumers have doubled."
And that's no idle reminiscence.  PolitiFact pretty much recycles the earlier fact check.  Rockefeller ended up with a "Half True" rating.  Van Hollen cherry-picked his years with greater care--2001 through 2006 instead of 2001 through 2009--so he fares a bit better.  Though I'm still wondering why PolitiFact does the comparison between years 2000 and 2006 while calling it the comparison between 2001 and 2006 (we won't see the same error when it comes to Claim B).

PolitiFact explains:
Back in 2009, we rated Rockefeller's claim Half True, noting that if he had carried the data forward to 2008 — an admittedly somewhat atypical year — the increase fell short of his stated quadrupling. But Van Hollen had a clearer reason to set the parameters between 2000 and 2006. Democrats regained control of the House in 2007. So it's fair for him to stop short of the 2008 figures.
If we're just verifying the numbers then the "clearer reason" that PolitiFact uses to rate Van Hollen higher than Rockefeller is irrelevant.  Apparently the underlying point has something to do with Democrats regaining control of the House in 2007.  Unfortunately, that doesn't entirely jibe with what Van Hollen said.

Van Hollen emphasized that Congress "did nothing" about the supposed problem of increased insurance company profits.  Unless Congress did something in 2007 then Van Hollen's "clearer reason" seems to have less to do with cause and effect than it does to fixing blame.  But blame for what?  Is it really a problem for insurance company profits to increase?

A parallel fact check by the Associated Press picked up on the importance of that issue.  Profit margins for the insurance industry remained relatively low during the period in question.  It turns out that measuring profit according to net profit reported by ten of the biggest insurers is a very misleading way to measure whether or not profits are reasonable.  A story from TheStreet.com helps illustrate:
The profit margin on commercial health insurance declined to 3.8% in 2006 on revenue of $172 billion from 4.9% on revenue of $164 billion in 2005.
Profits went up by Van Hollen's favored measure.  Profit margin went down.  And, at its best, the profit margin for health insurance companies comes in about average in the business world.  Liberal Washington Post columnist Ezra Klein put it this way:  "The insurance industry is not a particularly profitable industry."

Wasn't Van Hollen trying to make the opposite point?  If profits and premiums are both going up like crazy (quadrupling and doubling, respectively) then that means that the profits are excessive and perhaps Congress should intervene.  Isn't that the best and most sensible reading of Van Hollen's statement?

How did PolitiFact overlook that underlying point when it came time to rate Van Hollen?  The only thing keeping Van Hollen from a perfect "True" rating was apparently the fact that a good share of the rise in profits came from acquisitions--the top companies buying up smaller ones.  PolitiFact easily excused that bit of data massage:
Lawson acknowledged that the HCAN chart on profits does not address how much of an impact consolidation had on profits. But he added that the table actually comes from a study that specifically addressed the harm to competition caused by consolidation, so the group hardly ignored that issue. In addition, he defended the decision to use numbers unadjusted for consolidation. "That's what the market reacts to, and that's what they put in their annual report," Lawson said.
Those are not the droids we're looking for, in other words.

Investors' markets react to profits raised through acquisition.  And the reason for that is clear.  The larger company can then consolidate its operations and increase profits (and the efficient delivery of services) through the principle of economies of scale.  The same type of principle that liberals, by convenience, often turn into a huge positive when it comes to arguing for single-payer insurance.

Van Hollen chose a dubious measure of corporate profits (net profit as opposed to profit margin) and cherry-picked his group for measurement (a group of top insurers whose profits were enhanced by acquisitions). 

Claim B

PolitiFact:
Here we'll address Van Hollen's first claim that between 2000 and 2006 premiums in this country doubled. The time frame cited by Rockefeller is a little different, but the source for fact-checking it isn't.
Some say the underlying point is the most important thing when it comes to claims involving numbers.  I'm taking "doubled" as a number.
In 2000, average annual premiums for single people were $2,471, a number that rose to $4,242 by 2006 — a 72 percent increase. The amounts for family coverage rose over the same period from $6,438 to $11,480 — a 78 percent increase.
PolitiFact could have made the comparison to Van Hollen's claim more clear by describing the 2006 premium as 1.72 times the 2000 average premium for individuals and the 2006 premium as 1.78 times the 2000 premium for family plans.  Assuming that family plans outnumber individual plans, Van Hollen can squeak by on the accuracy of the figure.  Though we shouldn't forget that the underlying point is the most important thing.

Back to PolitiFact:
We also looked at just the portion of premiums paid by employees, because that's the number most consumers really care about. For individuals, the average premium paid by the employee went from $334 to $627; and for families, the employee contribution went from $1,619 to $2,973. That's an increase of 87 percent and 83 percent, respectively.

Those numbers get pretty close to Van Hollen's claim about doubling, but not quite. And so we rate his claim Mostly True.
Regardless of whether consumers "really care about" the out-of-pocket portion of their insurance premium, that has little to do with the insurance company.  Employers determine what portion of the insurance premium the employee will bear, not the company contracted to supply the insurance.  The final graph makes it appear that PolitiFact used the final set of numbers ("87 percent and 83 percent") in rating Van Hollen's statement.  The other set of numbers was more appropriate if we are to forget about things like profit margins and inflation.

Speaking of inflation, the inflation calculator at the Bureau of Labor statistics suggests that, adjusted for inflation, the 2006 premium is less than 1.5 times the 2000 premium for individual insurance.  For family insurance that drops the figure to a shade above 1.5 times the 2000 premium cost.

The latter portion of Van Hollen's claim exaggerates the increase of premiums, rounding up liberally from the applicable numbers.

Claims A and B

Altogether, Van Hollen argued that profits and insurance premiums were so high that Congress should have done something about it.  It's a terrible argument built on factoids containing a measure of truth, since profits and premiums have both increased considerably.  It's a terrible argument given the problems pointed out above.  But PolitiFact has little problem with it.

One supposes that the important underlying points consisted of "profits have increased" and "premiums have increased."  If that supposition is close to accurate, then we have a case where PolitiFact ignored the argument spelled out by the person making the claim.  Instead, PolitiFact picked an underlying point that might as well have been picked as a means of avoiding the clearly stated but problematic underlying point Van Hollen had in mind.

Profits went up largely, if not mostly, because the limited number of companies bought other companies and expanded their collective customer base.  Premiums went up because health costs went up while profit margins remained slim.  Van Hollen left out information that left the opposite impression from the one he intended to foment.  PolitiFact was either blind to it or did not care.


The grades:

Group A
Robert Farley:  F
Louis Jacobson:  F
Martha Hamilton: F

Group B
Robert Farley:  F
Louis Jacobson:  F
Martha Hamilton:  F

The twin fact checks lack due diligence at every turn.  If the underlying point is the most important aspect of a fact check regarding numbers then journalists err badly by missing the underlying point.

No comments:

Post a Comment

Please remain on topic and keep coarse language to an absolute minimum. Comments in a language other than English will be assumed off topic.