'Jobs Saved' Stats Are Worse Than Meaningless
Say what you will about the stimulus package, but it certainly has created one kind of job: a reporter who writes about how stimulus job creation numbers are bogus.
The federal government released a report on October 30 announcing that the $787 billion stimulus package had "created or saved" 640,000 jobs to date. The report has admirable specificity, reporting job creations and savings down to the employer and zip code. Trouble is, a lot of the data are wrong. Since the release, reporters at papers all over the country have busily written takedowns, picking out nonsensical factoids among the data.
The Chicago Tribune found that several Illinois school districts reported saving jobs for more teachers than they actually employ. The Financial Times learned that UCLA was counting tenured faculty among employees whose jobs were "saved" with stimulus dollars. Boston Globe reporters found a Head Start program in Greenfield, Mass., that claimed to save 90 jobs for $245,000 - an impressive $2,700 per employee. They also noted that the City of Revere created 64 jobs with a $485,500 project to install solar panels on a school's roof (presumably, that's 64 Green Jobs).
Moving to bigger fish, The Sacramento Bee noticed that the Cal State University system reported retaining 26,156 employees due to the stimulus - roughly half its workforce. Asked about this figure, which would seem to indicate a calamity averted, a CSU spokeswoman confirmed it was inflated and said, "This is not really a real number of people... it's like a budget number." (Only in California is "budget number" a synonym for "fake number.") It turned out stimulus funds were enough to pay half the system's employees for two months, not a whole year.
Some of these errors reflect confusion on the part of stimulus grantees when asked how many jobs they created, as is the case with a Kentucky shoe store owner who reported creating nine jobs with a contract for $889.60 - because he sold nine pairs of boots to the Army Corps of Engineers. But other errors come from the top. For example, the Department of Health and Human Services told grantees to reflect pay increases as fractional jobs created - as though a 5% pay increase is the same thing as creating 1/20 of a job.
And the errors add up. Reporters at the Washington Examiner reviewed the above reports and others from regional newspapers, concluding that at least 75,000 of the 640,000 "created or saved" jobs are bogus.
These vast data integrity problems are worthy of attention (and mockery). But by focusing on reporting issues, we risk giving the "jobs created or saved" statistic even more credit than it deserves. Even if all the reporting were fair and accurate, the stat would still be meaningless, because most of the jobs would have been saved anyway. Here's why. "Jobs created or saved" is a claim about the employment situation compared to a counterfactual world in which the stimulus was not enacted. Since this world doesn't exist, the government has to guess about what it would have been like. This is necessarily a subjective process with unverifiable results.
Unfortunately, the government has chosen a nonsensical counterfactual for the no-stimulus case. They've chosen to guess that every worker paid with stimulus dollars would have otherwise been laid off, as states worked to cut costs and close budget deficits. (Stimulus dollars are generally funneled through state and local budgets, and often used to pay for existing programs threatened by falling tax revenues.) Therefore, every job funded by the stimulus is a job created or saved.
This assumption doesn't accurately reflect the behavior of state and local governments. Layoffs and hiring freezes are only among the many measures governments use to close budget gaps. Alternatively, they can furlough workers. They can freeze or cut public employee salaries. They can raise taxes and fees. They can borrow money. They can sell assets. They can cut non-payroll expenses. In practice, states choose a combination of these and other methods to close budget gaps; they don't act solely by cutting jobs.
Numbers from the Bureau of Labor Statistics show that job cuts are not state and local governments' primary gap-closing measure. From July 2008 to July 2009, state and local government employment fell by 155,000 jobs. Assuming an average employment cost of $75,000 per worker, state governments saved about $12 billion through job cuts. But over that period, states closed budget gaps totaling $110 billion, meaning most of the gap closures came through other kinds of savings.
Indeed, this summer, Stateline.org reported that 21 states had furloughed or planned to furlough 728,500 employees. The Center on Budget and Policy Priorities writes that 30 states have raised taxes since the start of the recession. Actions like these aren't necessarily good policy, but they are seeing wide use as a way to avoid public sector job cuts.
Without stimulus funds, states and localities would have faced even bigger budget gaps, and they would have closed them in part by laying off more workers. But it's also safe to assume they would have found ways to retain most of the workers now funded by stimulus, just as they retained most of those who were threatened by budget gaps. The true stimulus job savings would only be a fraction of what's reported - even if the report accurately reflected the jobs funded by stimulus dollars.
This still leaves aside an even larger problem with "jobs created or saved" figures. They reflect "seen" effects - specific jobs funded by the stimulus - but not "unseen" ones. The figures cannot account for jobs lost because greater government expenditure leads to higher taxes or more government borrowing, discouraging private investment. On the other hand, they also don't account for jobs that might be spurred indirectly by stimulus-funded activity.
Ed DeSeve, a senior adviser to the President on the stimulus plan, defended the figures in to Wall Street Journal, noting that the program's purpose is "to create jobs, not count them." True. But the government releases count to show that the program is creating jobs. If the White House is hanging its hat on a fundamentally made-up number, why should we believe the stimulus is achieving its job creation purpose?
Newspapers like the Globe and the Bee have shown that the "jobs created or staved" emperor has no clothes. It's time for the White House to abandon this meaningless statistic and defend its fiscal policies on their merits.