Actually, there is no “Great Moments in Logic #1″ post, but I am mindful of a similarly themed post back in August 14, 2008, also motivated by a “Best of the Web” comment: “A True Canard? Thomas Friedman delghts via James Taranto.”
As a mathematician, displays of horrible logic always catch my eye, and Mr. Taranto has caught a whopper in a brief 12/1/2009 MarketWatch report by Rex Nutting. The quote below is taken from James Taranto’s daily “Best of the Web Today” feature in the online WSJ, which quotes the MarketWatch piece in full (one paragraph):
[From "Best of the Web"] – – – – – – – – – – – – – – – – – – – – – – – – – – – – – -
Stimuli Work, Ergo the Stimulus Is Working
MarketWatch reports on a cheering finding from the Congressional Budget Office:
The $787 billion fiscal stimulus program approved in February is working pretty much as expected, the Congressional Budget Office said Tuesday. U.S. employment is about 600,000 to 1.6 million higher, and real gross domestic product is about 1.2% to 3.2% higher than they would have been without the stimulus, CBO said. That estimate is nearly identical to the CBO’s assessment in March. The CBO based its estimates on long-standing relationships between additional spending and economic growth, rather than relying on incomplete and inaccurate reports from companies awarded federal contracts. Much of the direct government spending in the stimulus bill is yet to come.
Did we read that right? Yes, we did. Here’s the actual CBO report:
Estimating the law’s overall effects on employment requires a more comprehensive analysis than the recipients’ reports provide. Therefore, looking at the actual amounts spent so far (where identifiable) and estimates of the other effects of ARRA on spending and revenues, CBO has estimated the law’s impact on employment and economic output using evidence about how previous similar policies have affected the economy and various mathematical models that represent the workings of the economy. On that basis, CBO estimates that in the third quarter of calendar year 2009, an additional 600,000 to 1.6 million people were employed in the United States.
So the CBO’s estimate is “nearly identical” to the one in March because it is based on exactly the same information: the cost of the stimulus and the putative effect of spending that much money. It’s like a restaurant reviewer who estimates the quality of a meal by looking at the prices on the menu.
[End of Selection] – – – – – – – – – – – – – – – – – – – – – – – – – – – – – -
Did you catch that? Taranto summarized things wonderfully, but — as is my wont — let me overexplain…
There are a few things to point out: (1) The CBO’s statement that the stimulus “is working” is not based at all on any real look at the real results of the stimulus. It is essentially like saying, “Well, we projected in the past that it would work, and on the basis of those past projections we’ll say that it is working now.” Not that this means that the stimulus isn’t working, but that’s just the problem — it’s a meaningless statement. (2) Also, the MarketWatch paragraph, as so neatly highlighted by Taranto, says essentially that the CBO’s statement on Tuesday is the same as the CBO’s statement last March because it is based on exactly the same information — there is no new information, whatsoever. So why release a new statement, at all?
The report intial statement — “The $787 billion fiscal stimulus program approved in February is working pretty much as expected, the Congressional Budget Office said Tuesday.” — has, actually, no basis whatsoever. This would have been a more logically sound statement:
On Tuesday, the Congressional Budget Office reported its estimates of the cumulative effects of the $787 billion fiscal stimulus program approved in February. The CBO’s mathematical models predict that U.S. employment is 600,000 to 1.6 million jobs higher and the the real gross domestic product 1.2% to 3.2% higher than each would have been without the stimulus. The CBO reports that these figures are based on their model assumptions with considerations about where the stimulus money has been spent so far and not on actual measurements of real effects. Much of the direct government spending in the stimulus bill is yet to come.
It’s a shorter statement, too — normally, a difficult thing for me to achieve (see Exhibit A)!
As it stands, the reporting actually says, “The stimulus is working exactly as we expected, based solely on our assumption that it is working as expected.” And it’s not that the CBO report does not have potential value; it provides a basis by which the success/failure of the stimulus could be measured. But without any measurements, it’s useless. All of which makes the title of the MarketWatch piece — “Stimulus Working Pretty Much as Planned, CBO Says” — either a horribly inaccurate summary of what the CBO is actually saying or an inadvertent exposure to the CBO’s horrific reasoning on this point.
As usual, I don’t mean this as a comment on the politics or economics, just on the logic — or lack, thereof. And, I suppose, it is a comment on what bothers me about politics. “Data” like this is bandied about by anyone whose cause is best served, yet without any qualification or explanation of the information’s limitations. As part of an extended discussion, these projections could be helpful. As soundbites for campaign purposes or “points” scored on news programs, they are worthless.