As the cliche saying goes: there are lies, damn lies, and statistics. But the creation of the statistic for a job that is “saved or created” has got to be one of the worst statistical nightmares in history and it is completely meaningless, as it is entirely unprovable. Nevermind the accounting gaffes, the jobs created in zip codes and Congressional districts that don’t exist, and the accounts of dozens of jobs being saved for a few thousand dollars (would you take credit for saving a job that paid a person a few hundred dollars a year?). And nevermind that when the White House says “jobs” they really should be saying “job-years.” (Giving someone a job for one year is much different from creating a job that continues on in perpetuity.) The problem is that the statistic relies on both job creation by the spending itself, as well as jobs created by the “spending multiplier” — jobs created when the people whose job was ’saved or created” spend their money at their local stores, thereby “saving or creating” other jobs. In theory, you can do that. But in practice it’s near impossible, especially when those multipliers have clearly changed in our economy. People get concerned about the future and those people whose jobs were “saved or created” aren’t spending as much, so they’re not multiplying as much. And how do you account for the jobs that were lost because people stopped spending their money because of the expected increase in the budget deficit, or because of the fall in the dollar? If you’re going to go through a counter-factual scenario and do it in an honest way, you have to look at how everything changes, not just the things that go in your favor.
Some accounts put the cost of each of these jobs “saved or created” at over a half million dollars. The problem with that is that it treats all money as going to jobs, when some of it is being spent on infrastructure and capital equipment, which will provide benefits long after the stimulus is over.
Bottom line: it’s impossible to know how many jobs were saved or created. And touting a number, any number, when we’ve lost 7 million jobs in the recession just doesn’t hit home with people. That’s like me getting pulled over for going 30 mph over the speed limit in a school zone with kids everywhere and telling the cop, “At least I wasn’t going 40 mph over the limit! And I’m not even drinking either!” While it might be true, it probably won’t go far in convincing the cop to let me off.
Maybe I’m just being too skeptical. Maybe it really is easy to determine exactly how many jobs were “saved or created.” Or maybe not. This story from the Politico highlights the fact that three different Obama administration officials, on the same day, gave three different numbers for the number of jobs “saved or created” as a result of the stimulus bill:
(Good catch by Axelrod.) Any number they pick can never be proven and is based on a wealth of assumptions, but even still you’d think the least they could do is get together and come up with one number. Then maybe we’d believe them.
Today in his speech about the middle class, Obama claimed that the stimulus had ”saved or created” 2 million jobs, so that’s the number he’s going with. (Wise up, Gibbs!) Then he added that “economists agree,” even “conservative economists.” I find it fascinating that his own administration can’t agree on a number, but he is going to claim that economists of all stripes agree with the one specific number he gave today. Hogwash.
[On a side note, less than three months ago, at the end of December, the administration was saying that the stimulus bill had saved or created 1 million jobs. So for the last three months, when unemployment has been the highest in this recession, at 10%, and when we lost jobs every month but November, we've "saved or created" 1 million jobs -- as many as were "saved or created" in the previous 8 months. Do you believe that? Didn't think so.]
To me, the saddest part in all of this is that the “2 million” jobs “saved or created” is used as a defense that the stimulus worked. Tell that to the 7 million people who have lost jobs. The administration used to agree that “less bad” is not success:
“The first quarter of this year, we were losing jobs at an average of 700,000 jobs per month, month after month,” he said. “In the quarter that ended this week, the loss was 250,000 jobs per month, two-thirds less.”
Still, he said, “those facts and those realities aren’t good enough for President Obama, and they aren’t good enough for me.”
“We don’t think that ‘less bad’ is good,” Biden said. “‘Less bad’ is not our measure of success. One job lost is one job too many, and it’s still too much pain.”
If less bad is not your measure of success, why is it the first thing that comes out of the mouth of anyone from your administration whenever they are asked about the economy?
Using numbers from hypothetical situations is difficult, heavily dependent on assumptions that may never be proven, but the context in which you use these numbers is everything. Consider the following two scenarios:
a) You run a hospital that has had 1,000 staph infections in the previous year, so you institute a new protocol for hand-washing and this year you only have 200 staph infections. It’s pretty clear that the program is a success. And yes, even 1 staph infection is one too many, but you’ve made improvement and you can make a decent claim that your program made things better.
b) You run a hospital that had 1,000 staph infections the previous year, so you institute a new protocol and at the end of the year had 2,500 staph infections. Suppose you came out and said, “Yeah, staph infections were up this year, but we were hit by an epidemic and it would have been much worse without the measures we took. Without our protocol we would have had 4,000 staph infections.” Would people believe you? What if other reputable doctors were saying that the protocal wouldn’t have any impact (as some economists have said about the stimulus bill) and point to the increase in staph infections as proof that the protocol didn’t work?
Both of these examples rely on counterfactuals that can never be proven. But if you’re taking credit for something that cannot be proven at a time when the thing about which you are claiming success is going in the opposite direction, you’re probably going to have a hard time convincing people.
Footnote: This piece talks about how the task is so impossible that the administration is now not releasing employment updates because they can’t crunch the numbers. Even OMB director Peter Orszag has stopped using the phrase “saved or created” in favor of the more provable “funded by stimulus dollars” term. It’s going to make the number of jobs smaller, but at least this one is verifiable (assuming of course they can fix all the problems they have collecting the data).