Conn Carroll Can't Count

Conn Carroll has an amazingly duplicitous post over at the Washington Examiner, titled "Yes, Obamacare's Costs Have Almost Doubled." Here he is:

New York Magazine's Jonathan Chait, The New Republic's Jonathan Cohn, and The Washington Post's Ezra Klein all wrote posts have all written posts denying that Obamacare's costs have doubled since it became law. Below are the gross cost tables from both: 1) the March 18, 2010 CBO Obamacare cost estimate; and 2) the March 2012 Obamacare cost estimate.

$1.76 trillion is not double $940 billion, but it's close. Throw in one more year of full implementation ($265 billion in 2022) and the real ten year gross cost of Obamacare is north of $2 trillion.

Questions about the cost of government programs often involve time periods. Social Security, for example, has one cost over a single year and a different cost over five years, or 20 years. Looking at different numbers in this way allows us to think about how the costs are changing, and what we might do about it. If the cost went up, say, we could ask: is the cost structure fundamentally altering, or are more people entering the program? These are good things to think about.

However, when thinking in this way, it is important to be clear about when the program starts. This is the first boneheaded screwup Coll makes. Suppose I pass a law to start building a new jet fighter in 2014, and continue production for the next 20 years. Obviously, if I compare the window of 2010-2019 to the window of 2012-2021, the second will have a larger headline cost because it includes two more years of production. Obamacare is like this—most of the big programs start in 2014, as you see in the chart above when the spending increases by a factor of 5-10. Claiming this represents an increase in the cost of Obamacare is like saying a new aircraft carrier is cheap because its cost in 1960 was zero.

It gets worse, though. He hasn't even read his table. He quotes $1.76 trillion as the "gross cost," or the total spending ignoring revenues in the program, but his chart shows net costs of $1.25 trillion. Did he even glance at these tables before putting them up? He obviously didn't read the report, because $1.76 trillion is not in it anywhere. [UPDATE: A reader points out that I am mistaken here. That figure is in the table. Nevertheless, the remaining points stand.]

But that, unbelievably, isn't even the stupidest mistake Coll makes. (You might have spotted it already.) His tables don't have the same number of years. 2010-2019 is 10 years, which one can verify by counting if subtraction is too hard, and 2012-2022 is 11 years.

This is either transparent hack work, or straight-up idiotic. I'm not sure which is worse.

UPDATE: Ezra Klein adds graphs of the relevant years.